Muscle relaxants for pain management in rheumatoid arthritis (Review)

Transcription

1 Muscle relaxants for pain management in rheumatoid arthritis (Review) Richards BL, Whittle SL, Buchbinder R This is a reprint of a Cochrane review, prepared and maintained by The Cochrane Collaboration and published in The Cochrane Library 2012, Issue 1

2 T A B L E O F C O N T E N T S HEADER ABSTRACT PLAIN LANGUAGE SUMMARY SUMMARY OF FINDINGS FOR THE MAIN COMPARISON BACKGROUND OBJECTIVES METHODS RESULTS Figure Figure Figure Figure Figure Figure Figure Figure DISCUSSION AUTHORS CONCLUSIONS ACKNOWLEDGEMENTS REFERENCES CHARACTERISTICS OF STUDIES DATA AND ANALYSES Analysis 1.1. Comparison 1 Muscle relaxant versus control, Outcome 1 Pain 24hrs Analysis 1.2. Comparison 1 Muscle relaxant versus control, Outcome 2 Pain 1-2 weeks Analysis 2.1. Comparison 2 Benzodiazepine versus placebo, Outcome 1 Pain 24hrs Analysis 2.2. Comparison 2 Benzodiazepine versus placebo, Outcome 2 Pain 1 week Analysis 2.3. Comparison 2 Benzodiazepine versus placebo, Outcome 3 Sleep (MSLT) Analysis 2.4. Comparison 2 Benzodiazepine versus placebo, Outcome 4 Sleep (Polysomnography) Analysis 2.5. Comparison 2 Benzodiazepine versus placebo, Outcome 5 Sleep (Patient reported outcome measures).. 41 Analysis 2.6. Comparison 2 Benzodiazepine versus placebo, Outcome 6 Depression Analysis 3.1. Comparison 3 Benzodiazepine + NSAID versus NSAID - pain, Outcome 1 Pain 24hrs Analysis 3.2. Comparison 3 Benzodiazepine + NSAID versus NSAID - pain, Outcome 2 Sleep (Wolff Sleep Score).. 43 Analysis 4.1. Comparison 4 Non-benzodiazepine versus placebo, Outcome 1 Pain Analysis 4.2. Comparison 4 Non-benzodiazepine versus placebo, Outcome 2 Functional Status Analysis 4.3. Comparison 4 Non-benzodiazepine versus placebo, Outcome 3 Sleep (Polysomnography) Analysis 4.4. Comparison 4 Non-benzodiazepine versus placebo, Outcome 4 Sleep (Patient reported outcomes) Spiegel Sleep Questionnaire Analysis 4.5. Comparison 4 Non-benzodiazepine versus placebo, Outcome 5 Sleep (Patient reported outcomes) Leeds Sleep Evaluation Analysis 5.1. Comparison 5 Muscle relaxant versus control - safety, Outcome 1 Withdrawal due to adverse events.. 48 Analysis 5.2. Comparison 5 Muscle relaxant versus control - safety, Outcome 2 Total Adverse Events Analysis 5.3. Comparison 5 Muscle relaxant versus control - safety, Outcome 3 Total Adverse events - trials greater than 24hrs duration Analysis 5.4. Comparison 5 Muscle relaxant versus control - safety, Outcome 4 Total adverse events - trials 24hr duration only Analysis 5.5. Comparison 5 Muscle relaxant versus control - safety, Outcome 5 Subgroups Adverse Events APPENDICES WHAT S NEW HISTORY CONTRIBUTIONS OF AUTHORS DECLARATIONS OF INTEREST i

3 SOURCES OF SUPPORT ii

4 [Intervention Review] Muscle relaxants for pain management in rheumatoid arthritis Bethan L Richards 1, Samuel L Whittle 2, Rachelle Buchbinder 3 1 Institute of Rheumatology and Orthopedics, Royal Prince Alfred Hospital, Camperdown, Australia. 2 Rheumatology Unit, The Queen Elizabeth Hospital, Woodville, Australia. 3 Monash Department of Clinical Epidemiology at Cabrini Hospital, Department of Epidemiology and Preventive Medicine, School of Public Health and Preventive Medicine, Monash University, Malvern, Australia Contact address: Bethan L Richards, Institute of Rheumatology and Orthopedics, Royal Prince Alfred Hospital, Missenden Road, Camperdown, New South Wales, 2050, Australia. brichard@med.usyd.edu.au. Editorial group: Cochrane Musculoskeletal Group. Publication status and date: New, published in Issue 1, Review content assessed as up-to-date: 6 September Citation: Richards BL, Whittle SL, Buchbinder R. Muscle relaxants for pain management in rheumatoid arthritis. Cochrane Database of Systematic Reviews 2012, Issue 1. Art. No.: CD DOI: / CD pub2. Background A B S T R A C T Pain management is a high priority for patients with rheumatoid arthritis (RA). Muscle relaxants include drugs that reduce muscle spasm (for example benzodiazepines such as diazepam (Valium), alprazolam (Xanax), lorazepam (Ativan) and non-benzodiazepines such as metaxalone (Skelaxin) or a combination of paracetamol and orphenadrine (Muscol)) and drugs that prevent increased muscle tone (baclofen and dantrolene). Despite a paucity of evidence supporting their use, antispasmodic and antispasticity muscle relaxants have gained widespread clinical acceptance as adjuvants in the management of patients with chronic musculoskeletal pain. Objectives The aim of this review was to determine the efficacy and safety of muscle relaxants in pain management in patients with RA. The muscle relaxants that were included in this review are the antispasmodic benzodiazepines (alprazolam, bromazepam, chlordiazepoxide,cinolazepam, clonazepam, cloxazolam, clorazepate, diazepam, estazolam, flunitrazepam, flurazepam, flutoprazepam, halazepam, ketazolam, loprazolam, lorazepam, lormetazepam, medazepam, midazolam, nimetazepam, nitrazepam, nordazepam, oxazepam, pinazepam, prazepam, quazepam, temazepam, tetrazepam, triazolam), antispasmodic non-benzodiazepines (cyclobenzaprine, carisoprodol, chlorzoxazone, meprobamate, methocarbamol, metaxalone, orphenadrine, tizanidine and zopiclone), and antispasticity drugs (baclofen and dantrolene sodium). Search methods We performed a search of the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library, 4th quarter 2010), MEDLINE (1950 to week 1 November 2010), EMBASE (Week ), and PsycINFO (1806 to week 2 November 2010). We also searched the 2008 to 2009 American College of Rheumatology (ACR) and European League Against Rheumatism (EULAR) abstracts and performed a handsearch of reference lists of relevant articles. Selection criteria We included randomised controlled trials which compared a muscle relaxant to another therapy (active, including non-pharmacological therapies, or placebo) in adult patients with RA and that reported at least one clinically relevant outcome. Data collection and analysis Two blinded review authors independently extracted data and assessed the risk of bias in the trials. Meta-analyses were used to examine the efficacy of muscle relaxants on pain, depression, sleep and function, as well as their safety. 1

5 Main results Six trials (126 participants) were included in this review. All trials were rated at high risk of bias. Five cross-over trials evaluated a benzodiazepine, four assessed diazepam (n = 71) and one assessed triazolam (n = 15). The sixth trial assessed zopiclone (a nonbenzodiazepine) (n = 40) and was a parallel group study. No trial duration was longer than two weeks while three single dose trials assessed outcomes at 24 hours only. Overall the included trials failed to find evidence of a beneficial effect of muscle relaxants over placebo, alone (at 24 hrs, 1 or 2 weeks) or in addition to non-steroidal anti-inflammatory drugs (NSAIDs) (at 24 hrs), on pain intensity, function, or quality of life. Data from two trials of longer than 24 hours duration (n = 74) (diazepam and zopiclone) found that participants who received a muscle relaxant had significantly more adverse events compared with those who received placebo (number needed to harm (NNTH) 3, 95% CI 2 to 7). These were predominantly central nervous system side effects, including dizziness and drowsiness (NNTH 3, 95% CI 2 to 11). Authors conclusions Based upon the currently available evidence in patients with RA, benzodiazepines (diazepam and triazolam) do not appear to be beneficial in improving pain over 24 hours or one week. The non-benzodiazepine agent zopiclone also did not significantly reduce pain over two weeks. However, even short term muscle relaxant use (24 hours to 2 weeks) is associated with significant adverse events, predominantly drowsiness and dizziness. P L A I N L A N G U A G E S U M M A R Y Muscle relaxants for pain management in rheumatoid arthritis This summary of a Cochrane review presents what we know from research about the effect of muscle relaxants on pain in patients with rheumatoid arthritis. The review shows that in people with rheumatoid arthritis: - Muscle relaxants may not improve pain when taken as a single dose or for up to a two week period - We are uncertain whether muscle relaxants affect functional status because of the very low quality of the evidence - No trials were found that evaluated whether muscle relaxants affect quality of life - No trials were found that evaluated whether antidepressants affect sleep - We are uncertain whether muscle relaxants affect mood because of the very low quality of the evidence We also do not have precise information about side effects and complications. This is particularly true for rare but serious side effects. Possible side effects may include feeling tired or nauseous, headaches, blurred vision, a dry mouth, sexual dysfunction, or becoming dizzy or constipated. Rare complications may include increased suicidal thinking, liver inflammation, or a reduced white cell count. What is rheumatoid arthritis and what are muscle relaxants? When you have rheumatoid arthritis your immune system, which normally fights infection, attacks the lining of your joints. This makes your joints swollen, stiff, and painful. The small joints of your hands and feet are usually affected first. There is no cure for rheumatoid arthritis at present, so the treatments aim to relieve pain and stiffness and improve your ability to move. Muscle relaxants can be used to treat anxiety and promote sleep, and some people believe they may also reduce pain by acting on the nerves that cause pain, but this remains controversial. Muscle relaxants include drugs that reduce muscle spasm (for example benzodiazepines such as diazepam (Valium), Xanax, Ativan and non-benzodiazepines such as Skelaxin, Muscol) and drugs that prevent increased muscle tone (baclofen and dantrolene). Best estimates of what happens to people with rheumatoid arthritis who take muscle relaxants: Pain at 24 hours: - Non-significant result. Pain at 1 to 2 weeks: - Non-significant result. 2

6 Withdrawal due to adverse events, after 2 weeks: - Non-significant result. Total adverse events: - 49 more people out of 100 experienced an adverse event, after 1 to 2 weeks, when they took a muscle relaxant (absolute difference 49%), - 52 out of 100 people who took a muscle relaxant suffered an adverse event, - 3 out of 100 people who took a placebo suffered an adverse event. Central nervous system (CNS) adverse events: - 33 more people out of 100 experienced a CNS adverse event, after 1 to 2 weeks, when they took a muscle relaxant (absolute difference 33%), - 39 out of 100 people who took a muscle relaxant suffered a CNS adverse event, - 6 out of 100 people who took a placebo relaxant suffered a CNS adverse event. This record should be cited as: This is a Cochrane review abstract and plain language summary, prepared and maintained by The Cochrane Collaboration, currently published in the Cochrane Database of Systematic Reviews [Issue and date] [year] The Cochrane Collaboration. Published by John Wiley and Sons, Ltd.. The full text of the review is available in The Cochrane Library (ISSN X). 3

7 S U M M A R Y O F F I N D I N G S F O R T H E M A I N C O M P A R I S O N [Explanation] Muscle relaxant versus control for pain management in rheumatoid arthritis Patient or population: patients with pain management in rheumatoid arthritis Settings: Intervention: muscle relaxant versus control Outcomes Illustrative comparative risks*(95% CI) Relative effect (95% CI) Pain - 24hrs (Single dose) Follow-up: 24 hours Assumed risk Control Corresponding risk Muscle Relaxant versus control The mean Pain - 24hrs (Single dose) in the intervention groups was 0.22 lower (1.02 lower to 0.58 higher) No of Participants (studies) 104 (3 studies) Quality of the evidence (GRADE) low 1,2 Comments No significant difference. Absolute risk difference 2%(6%to10%)andrelative percent change 8%(- 38%to22%) Pain-1-2weeks Follow-up: 1 weeks The mean Pain weeks in the intervention groups was 0.20 standard deviations lower (0.59 lower to 0.18 higher) 104 (3 studies) verylow 2,3,4 No significant difference. Absolute risk difference - 5%(-15%to5%)andrelative percent change-4% (-11%to3%) Withdrawal due to Adverse Events Follow-up: 2 weeks 0per1000 0per1000 (0to0) RR 2.84 (0.31 to 26.08) 180 (5studies 5 ) verylow 2,3,4 Theeventrateinthecontrol group was zero. Absolute risk difference 1% (-4% to 6%), relative percent change 184%(-69% to251%) 4

8 Total Number of Adverse Events - only studies >24hrs duration Follow-up: 2 weeks Total Number of Adverse Events - single dose studies Follow-up: 24 hours CNS adverse events Follow-up: 1-2 weeks 29per per per per1000 (31to438) 403per1000 (167to982) 340per1000 (101 to 1000) RR 4.03 (1.08 to 15.10) RR 1.40 (0.58 to 3.41) RR 5.96 (1.77 to 20.08) 74 (2 studies) 106 (3 studies) 74 (2 studies) verylow 2,3,4 verylow 2,3 verylow 2,3,4 Number needed to harm (NNTH) was 3 (2 to 8). Absolute risk difference 40% (23% to 57%), relative percentage change 303%(8%to1410%) No significant difference. Absolute risk difference - 6% (-23% to 10%) and relative percent change - 22%(-59%to48%) NNTH was3(ci2to11). Absolute risk difference 35%(-13%to83%),relative percent change 496% (77% to 1908%) *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention(and its 95% CI). CI: Confidence interval; RR: Risk ratio; GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Lowquality:Furtherresearchisverylikelytohaveanimportantimpactonourconfidenceintheestimateofeffectandislikelytochangetheestimate. Very low quality: We are very uncertain about the estimate. 1 Hetergeneousinterventionsandcontrols 2 Wideconfidenceintervalsgivensmallnumberofparticipantsandsmalleventrate 3 Allthreestudieshadhighriskofbias. 4 Heterogenousinterventions,outcomes,studydesignandlengthoffollowup 5 Threeofthefivetrialshandeventratesof0 5

9 B A C K G R O U N D Description of the condition Rheumatoid arthritis (RA) is the most common form of inflammatory arthritis, affecting around 1% of the population. However, despite the significant advances in treatment over the past few decades, pain management remains a significant issue for many patients (Heiberg 2002). Progressive disease is characterized by synovial tissue proliferation and persistent inflammation with resultant cartilage degradation, bone erosion, and damage to adjacent soft tissue and neural structures (Tak 2000). In the early stages of RA, pain reflects the nociceptive effects of local inflammation (Kidd 2001). Over time, however, the sources of pain are more numerous (Bolay 2002) and the pain is often compounded by associated poor sleep, psychological comorbidity, and muscle tension (Wolfe 2006). Hypnotic agents or muscle relaxants are widely prescribed in the management of insomnia, anxiety, or muscle tension. In a survey of 242 patients with RA, 60% reported that arthritis pain interfered with sleep to some degree, with an additional 14% reporting severe or very severe interference (Nicassio 1992). When a patient is unable to sleep because of pain, a sleep-promoting agent may be appropriate as increasing sleep may help to increase daytime pain thresholds. Poor sleep, independent of mood status, has also been associated with fatigue, exhaustion, irritability, poor function, and a cycle of greater pain (Morin 1998). This may also contribute to muscle tension, which has also been linked to increased pain in RA ( Koehler 1985) and osteoarthritis (OA) (Dekker 1992). There is also literature to support the notion that elevated levels of anxiety are seen in patients with RA, also associated with higher levels of pain (Dickens 2002; Hagland 1989; Smedstad 1996; Smedstad 1997). Muscle relaxants may therefore be useful adjuvant agents in patients with RA who have persistent pain. Description of the intervention The term muscle relaxants is very broad and includes a wide range of drugs with different indications and mechanisms of action. Muscle relaxants can be divided into two main categories, antispasmodic and antispasticity medications. The antispasmodic agents are further subclassified into the benzodiazepines and the non-benzodiazepines. Since the introduction of chlordiazepoxide (Librium ) in 1960 (Tobin 1960) and diazepam (Valium ) in 1962 (Randall 1961), the benzodiazepines have been widely prescribed for a variety of medical and psychiatric indications. Non-benzodiazepines include a variety of drugs that can act at the brain stem or spinal cord level. These include carisoprodol, chlorzoxazone, cyclobenzaprine, metaxalone, meprobamate, methocarbamol, tizanidine, zopiclone, and orphenadrine. While these agents have been used for the treatment of painful musculoskeletal conditions that are associated with muscle spasm such as acute low back pain and muscle strains (Waldman 1994), their use in RA is less well described. Although these drugs may relieve skeletal muscle pain, their effects are non-specific and not solely related to muscle relaxation. They exhibit modest analgesic activity that may derive from their sedative effects and also possibly from suppression of nociceptive input (Hunskaar 1991). All antispasmodic agents can cause significant drowsiness, dizziness, confusion, nausea, and vomiting. Antispasticity medications are used to reduce spasticity that interferes with therapy or function. Examples of such agents include baclofen and dantrolene. They are not commonly used in patients with RA. How the intervention might work In recent years we have witnessed dramatic advances in our understanding of the neuroanatomy of pain pathways and the neurophysiology of pain regulation. In RA, pain frequently has multifactorial origins with both central and peripheral components. Pain thresholds are known to be modified by sleep, however few current neurobiological hypotheses adequately explain the complex relationship between chronic pain and sleep disturbance. To date, sleep and pain are known to use common neurotransmitters (Moldofsky 1975) and alpha wave intrusion into non-rapid eye movement (REM) sleep has been suggested as a possible mechanism of sleep disturbance in patients with fibromyalgia (Branco 1994). Molecular mechanisms linking psychological state and pain have also been recognized. Patients with pain often report increases in pain in conjunction with elevations in psychological stress and the molecular link between psychological stress and pain may be explained by the stress-induced increases in tumour necrosis factor (TNF) and interleukin (IL)-1 (Maes 1998). Centrally, inhibitory gamma-amino butyric acid (GABA) neurons in the spinal cord are known to act as gate keepers and to control the relay of pain signals from the periphery to higher areas of the central nervous system. This pivotal role of GABA neurons in modulating pain signals has been demonstrated in several reports which have shown that a loss of such inhibitory capabilities underlies several forms of chronic pain (Malan 2002; Moore 2002). In addition, a recent animal study showed that specifically activating spinal GABA A receptors containing α2 or α3 subunits reduced nociceptive input and emotional processing of inflammatory and neuropathic pain (Knabl 2008). Peripheral benzodiazepine receptors (PBR) have also been shown to be involved in the regulation of immune responses (Waterfield 1999) and PBR ligands exhibit anti-inflammatory properties (Zavala 1990). The analgesic effects of benzodiazepines are predominantly mediated through activation of neuronal GABA A receptors (Costa 1979), although benzodiazepines may also act via peripheral mechanisms in patients with RA. The non-benzodiazepines, however, are structurally unrelated compounds that may indirectly relax 6

10 skeletal muscle by blocking postsynaptic neurons in the spinal cord and the descending reticular formation in the brain. Baclofen is a gamma aminobutyric acid (GABA) derivative that inhibits neural transmission at the spinal level and also depresses the central nervous system. Dantrolene sodium blocks the sarcoplasmic reticulum calcium channel thereby diminishing the actin-myosin interaction, causing muscle relaxation. Why it is important to do this review Since their introduction, there has been interest in the therapeutic application of muscle relaxants for the management of pain. Conclusive data regarding their analgesic activity, however, is lacking. There is also conflicting evidence as to whether or not muscle relaxants possess analgesic properties that are independent of their effects on sleep. Despite a paucity of evidence to support their use, muscle relaxants have gained widespread clinical acceptance as adjuvants in the management of patients with chronic musculoskeletal pain (Gordon 1995; Levy 1994). They have also been advocated for pain associated with anxiety (Fernandez 1987), muscle injury (Lossius 1980), muscle spasm (Weber 1973), and nerve injury (Smirne 1977). This review helps to clarify the risks and benefits associated with using muscle relaxants in the management of pain in patients with RA to aid physicians in making a more informed decision about their use. O B J E C T I V E S The objectives of this review were to evaluate the analgesic efficacy and safety of muscle relaxants in patients with RA. M E T H O D S Criteria for considering studies for this review Types of studies All published randomised (RCTs) or quasi-randomised (that is where allocation was not truly random) (CCTs) controlled trials which compared muscle relaxant therapy to another therapy (active, including non-pharmacological therapies, or placebo) for RA were considered for inclusion. Only trials that were published as full articles or were available as a full trial report were included. Types of participants Adult patients (aged 18 years or older) with a diagnosis of RA. Populations that included a mixed population with RA and other causes of musculoskeletal pain were excluded unless results for the RA population could be separated out from the analysis. Types of interventions All formulations and doses of muscle relaxants, as monotherapy or in combination, were considered. Comparators could include: 1. placebo; 2. other analgesics (e.g. paracetamol, non-steroidal antiinflammatory drugs (NSAIDs), opioids, tramadol, neuromodulators etc); 3. non-pharmacological modalities (e.g. transcutaneous electrical nerve stimulation (TENS), acupuncture, etc); 4. same drug at differing doses; and 5. other muscle relaxants. Comparisons with placebo and with other controls were planned to be reported separately. Drugs that had been withdrawn from the market due to safety concerns were excluded from the review. Types of outcome measures There is considerable variation in the outcome measures reported in clinical trials of interventions for pain. For the purpose of this systematic review, we included outcome measures recommended by the Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials (IMMPACT) and Dworkin As continuous outcome measures in pain trials (such as mean change on a 100 mm visual analogue scale) may not follow a Gaussian distribution, and often a bimodal distribution is seen instead (where patients tend to report either very good or very poor pain relief) (Moore 2010), there is a difficulty in interpreting the meaning of average changes in continuous pain measures. For this reason, a dichotomous outcome measure (the proportion of participants reporting 30% pain relief) was likely to be more clinically relevant and was the primary efficacy measure in this review. It is recognised, however, that it has been the practice in most trials of interventions for chronic pain to report continuous measures and therefore the mean change in pain score was also included as a secondary efficacy measure. A global rating of treatment satisfaction, such as the Patient Global Impression of Change scale (PGIC), which provides an outcome measure that integrates pain relief, changes in function, and side-effects into a single, interpretable measure, is also recommended by IMMPACT and was included as a secondary outcome measure (Dworkin 2008). Main outcomes 1. Efficacy: patient reported pain relief of 30% or greater. 7

11 2. Safety: number of withdrawals due to adverse events. Minor outcomes 3. Pain: a. patient reported pain relief of 50% or greater; b. patient reported global impression of clinical change (PGIC) much or very much improved; c. proportion of patients achieving pain score below 30/100 mm on a visual analogue scale; or d. mean change in pain score on a visual analogue scale or numerical rating scale. 4. Number and types of adverse events (AEs) and serious adverse events (SAEs) (defined as AEs that were fatal, life-threatening, or required hospitalisation). 5. Function: as measured by the Health Assessment Questionaire (HAQ) or modified HAQ (Fries 1980; Pincus 1983). 6. Quality of life: as measured by either generic instruments (such as the Short Form-36 (SF-36)) or disease-specific tools (such as the Rheumatoid Arthritis Quality of Life instrument (RAQoL)). 7. Participant withdrawals due to inadequate analgesia. 8. Sleep as measured by any commonly used sleep scale (eg. Insomnia Severity Index, Medical Outcomes Study (MOS), Sleep Scale, Pittsburgh Sleep Diary (PSD), and Pittsburgh Sleep Quality Index (PSQI)). 9. Depression as measured by any commonly used depression scale (eg. Hamilton Rating Scale for Depression (HRSD), Hospital Anxiety and Depression (HAD) score, Beck Depression Inventory (BDI), Zung self rating depression score). The duration of the trials of interventions for pain varies considerably. The efficacy of interventions, and the relative balance of benefits and harms, may vary according to the duration of the trial and therefore the combination of results from trials of different durations may represent a source of bias in systematic reviews (Moore 2010). For the purpose of this review, trials were grouped into those of duration < 1 week, 1 to 6 weeks, and > 6 weeks. Where available, the short and long term outcomes for the proportion reporting pain relief of 30% or greater, total number of withdrawals due to adverse effects, number of serious adverse events, function, and quality of life were presented in the summary of findings table. Search methods for identification of studies Electronic searches To identify relevant trials for this review, we used computer-aided searches of the following databases for RCTs or CCTs using the search strategies detailed in the appendices: 1. Ovid MEDLINE (1950 to week 1 November 2010) (Appendix 1); 2. EMBASE Classic + EMBASE (Week ) (Appendix 2); 3. Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library, 4th quarter 2010); 4. PsycINFO (1806 to week 2 November 2010). No language restrictions were applied. Searching other resources The American College of Rheumatology (ACR) and European League Against Rheumatism (EULAR) conference abstracts from 2008 and 2009 were searched. Handsearches of references and relevant reviews were also performed to identify any additional trials not retrieved by the above methods. Data collection and analysis Selection of studies All identified studies were assessed independently by two review authors (BR and SW) to identify trials that fulfilled the inclusion criteria. All possibly relevant articles were retrieved in full text and any disagreement in study selection was resolved by consensus or by discussion with a third review author (RB). Data extraction and management Two independent review authors (BR and SW) extracted relevant information from the included trials including study design, characteristics of study population, treatment regimen and duration, outcomes and timing of outcome assessment using predetermined forms. The raw data (means and standard deviations for continuous outcomes and number of events or participants for dichotomous outcomes) were extracted for outcomes of interest. s in data extraction were resolved by referring back to the original articles and establishing consensus. A third review author (RB) was consulted to help resolve differences, as necessary. Assessment of risk of bias in included studies Two authors (BR, SW) independently assessed risk of bias for all included studies for the following items: random sequence generation; allocation concealment; blinding of participants, care provider, and outcome assessor for each outcome measure (see primary and secondary outcome measures); incomplete outcome data; and other biases, conforming to the methods recommended by The Cochrane Collaboration (Higgins 2008). To determine the risk of bias of a study, for each criterion the presence of sufficient information and the likelihood of potential bias were evaluated. Each criterion was rated as yes (low risk of bias), no (high risk of bias), or unclear (either lack of information or uncertainty over the potential for bias). In a consensus meeting, disagreements 8

12 among the review authors were discussed and resolved. If consensus could not be reached, a third review author (RB) made the final decision. Measures of treatment effect The data were summarised in a meta-analysis only if there was sufficient clinical and statistical homogeneity. For continuous data, results were analysed as mean differences between the intervention and comparator groups (MD) with 95% confidence intervals. The mean difference between the treated group and the control group was weighted by the inverse of the variance in the pooled treatment estimate. However, when different scales were used to measure the same conceptual outcome (for example functional status or pain), standardized mean differences (SMD) were calculated instead. SMDs were calculated by dividing the MD by the standard deviation, resulting in a unitless measure of treatment effect. For dichotomous data, a relative risk (RR) with corresponding 95% confidence interval was calculated. Unit of analysis issues For studies containing more than two intervention groups, making multiple pair-wise comparisons between all possible pairs of intervention groups possible included the same group of participants only once in the meta-analysis. Cross-over trials were identified, in which the reporting of continuous outcome data precluded paired analysis, however there was no meta-analysis so unit-of-analysis error was not an issue. Dealing with missing data In cases where individuals were missing from the reported results, we assumed the missing values to have a poor outcome. For dichotomous outcomes that measured adverse events (for example number of withdrawals due to adverse events), the withdrawal rate was calculated using the number of patients that received treatment as the denominator (worst case analysis). For dichotomous outcomes that measured benefits (for example proportion of participants achieving an American College of Rheumatology 20% improvement criteria (ACR20) response) the worst case analysis was calculated using the number of randomised participants as the denominator. For continuous outcomes (for example pain) we calculated the MD or SMD based on the number of patients analysed at the time point. If the number of patients analysed was not presented for each time point, the number of randomised patients in each group at baseline was used. Where possible, missing standard deviations were computed from other statistics such as standard errors, confidence intervals or P values according to the methods recommended in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2009). Assessment of heterogeneity Prior to meta-analysis, we assessed studies for clinical homogeneity with respect to type of therapy, control group, and the outcomes. For any studies judged as clinically homogeneous, statistical heterogeneity was estimated using the I 2 statistic (Deeks 2008) with the following as a rough guide for interpretation: 0% to 40% might not be important, 30% to 60% may represent moderate heterogeneity, 50% to 90% may represent substantial heterogeneity, and 75% to 100% may represent considerable heterogeneity. Assessment of reporting biases In order to determine whether reporting bias was present, we planned to determine whether the protocol of the RCT was published before recruitment of study patients was started. However as no studies were published after 1st July 2005, we did not carry out the preplanned screen of the Clinical Trial Register at the International Clinical Trials Registry Platform of the World Health Organization ( (DeAngelis 2004). We also evaluated whether selective reporting of outcomes was present (outcome reporting bias). We compared the fixed-effect model estimate against the randomeffects model to assess the possible presence of small sample bias in the published literature (that is in which the intervention effect was more beneficial in smaller studies). In the presence of small sample bias, the random-effects model estimate was used (Sterne 2008). The potential for reporting bias was planned to be further explored by funnel plots if 10 studies were available, however due to the limited number of studies identified this was not done. Data synthesis Where studies were sufficiently homogeneous that it remained clinically meaningful for them to be pooled, meta-analysis was performed using a random-effects model, regardless of the I 2 statistic results. Analysis was performed using Review Manager 5 and forest plots were produced for all analyses. Subgroup analysis and investigation of heterogeneity If sufficient data had been available, the following subgroup analyses were planned: 1. patients ages (< 65 years versus 65 years); 2. gender (male versus female); and 3. duration of RA ( 2 years versus > 2 years). Sensitivity analysis If sufficient data had been available, we planned sensitivity analyses to assess the impact of any bias attributable to inclusion of trials with inadequate treatment allocation concealment (including studies with quasi-randomised designs). 9

13 Presentation of key results A summary of findings table was produced using GRADEpro software. This table provides key information concerning the quality of evidence, the magnitude of effect of the interventions examined, and the sum of available data on the outcomes (short and long term outcomes for pain, total number of withdrawals due to adverse effects, function, and quality of life), as recommended by The Cochrane Collaboration (Schünemann 2008a). The table includes an overall grading of the evidence related to each of the main outcomes, using the GRADE approach. In addition to the absolute and relative magnitudes of effect provided in the summary of findings table, for dichotomous outcomes the number needed to treat to benefit (NNTB) or the number needed to treat to harm (NNTH) was calculated from the control group event rate (unless the population event rate was known) and the relative risk was calculated using the Visual Rx NNT calculator (Cates 2004). For continuous outcomes, the NNT was calculated using the Wells calculator software, available at the Cochrane Musculoskeletal Group (CMSG) editorial office ( The minimal clinically important difference (MCID) for each outcome was determined for input into the calculator. R E S U L T S Description of studies See: Characteristics of included studies; Characteristics of excluded studies. See: Characteristics of included studies ; Characteristics of excluded studies. Results of the search The database search yielded a total of 174 articles for review (CEN- TRAL 26, MEDLINE 95, EMBASE 49, and PsycINFO 4) and no further relevant studies were identified from searching the 2008 and 2009 ACR and EULAR abstracts. After removal of 42 duplicates, the records were screened and nine studies were assessed for detailed review. Six trials (n = 126 participants) met the inclusion criteria of the review (Bayley 1976; Drewes 1998; Hobkirk 1977; Sharma 1978; Vince 1973; Walsh 1996). No additional studies were identified through reference checking (Figure 1). 10

14 Figure 1. Study flow diagram. 11

15 Included studies The characteristics of the included studies are described in the Characteristics of included studies table. Five trials assessed a benzodiazepine and one trial assessed the non-benzodiazepine zopiclone (n = 40) (Drewes 1998). Of the benzodiazepine studies, four evaluated diazepam (n = 71) (Bayley 1976; Hobkirk 1977; Sharma 1978; Vince 1973) and one evaluated triazolam (n = 15) (Walsh 1996). Five trials included a placebo control (Bayley 1976; Drewes 1998; Hobkirk 1977; Vince 1973; Walsh 1996), one compared diazepam with an NSAID (Bayley 1976), and two studies assessed whether diazepam in combination with an NSAID was superior to an NSAID alone (Hobkirk 1977; Sharma 1978). A summary of the interventions studied in the six included trials is listed below. 1) Benzodiazepine versus placebo 1a) Diazepam versus placebo (Bayley 1976; Vince 1973) 1b) Triazolam versus placebo (Walsh 1996) 2) Benzodiazepine versus NSAID 2a) Diazepam versus Indomethacin (Bayley 1976) Five trials used a cross-over design (Bayley 1976; Hobkirk 1977; Sharma 1978; Vince 1973; Walsh 1996) and, of these, only one included a washout period (Walsh 1996). The remaining study used a parallel group design (Drewes 1998). No benzodiazepine trial incorporated more than one week of active treatment and, overall, no trial was longer than two weeks; the shortest duration trials were three cross-over studies of single doses of drug given for three consecutive nights (Bayley 1976; Hobkirk 1977; Sharma 1978). The other two cross-over studies included two one week periods of treatment (Vince 1973; Walsh 1996). Three studies evaluated inpatients (Bayley 1976; Hobkirk 1977; Sharma 1978) and three studies (Drewes 1998; Vince 1973; Walsh 1996) included outpatients. Most participants were women (83%), in accordance with the epidemiology of RA, and all patients had active disease with 56% of patients hospitalised at the time of study. One trial incorporated patients who had both RA and sleep impairment (Walsh 1996). Only 17% (22/127) of patients were receiving corticosteroids and 30% (38/127) were receiving disease modifying antirheumatic drugs (DMARDs) (Walsh 1996) (with only 5/38 on methotrexate). No patients were receiving biological agents. No studies reported any specific information about the type of pain participants were suffering from or whether patients suffered from depression. 3) Benzodiazepine + NSAID versus NSAID 3a) Diazepam + sulindac versus sulindac (Sharma 1978) 3b) Diazepam + indomethacin versus indomethacin (Hobkirk 1977) 4) Non-benzodiazepine versus placebo 4a) Zopliclone versus placebo (Drewes 1998) The majority of studies were published in the late 1970s, with the most recent publications being Walsh 1996 and Drewes Excluded studies Three studies were excluded from this review (see Characteristics of excluded studies table) because they included mixed populations and it was not possible to extract data regarding the RA patients alone for analysis (Durrigl 1969; Hardo 1991; Tarpley 1965). Risk of bias in included studies See Figure 2. 12

16 Figure 2. Risk of bias summary: review authors judgements about each risk of bias item for each included study. 13

17 All studies were rated at high risk of bias (Bayley 1976; Drewes 1998; Hobkirk 1977; Sharma 1978; Vince 1973; Walsh 1996). The most common methodological shortcomings in the studies involved the following, in order of frequency. Inadequate method of randomisation (6, 100% trials scored unclear ). Inadequate concealment of the drug allocation procedures (6, 100% trials scored unclear ). Failing to apply intention-to-treat analysis (5, 87% trials scored negative or unclear ). Non-equivalent co-interventions (4, 67% scored unclear ). Dissimilarity of the baseline characteristics (4, 67% trials scored negative or unclear ). Failure to blind study personnel (3, 50% trials scored unclear ). Failing to evaluate compliance (3, 50% trials scored negative or unclear ). Inadequate dropouts (2, 33% trials scored negative or unclear ). Failure to address incomplete outcome data (1, 17% trials scored negative or unclear ). Failure to blind participants (1, 17% trials scored negative or unclear ). Only reporting selective outcomes (no trials scored negative or unclear ). Allocation No study adequately described how participants were randomised and all were deemed unclear. No study provided any information about whether allocation of treatment was adequately concealed and all were deemed to be unclear. Incomplete outcome data Only one study failed to describe how incomplete outcome data were addressed. In Vince 1973, 7/24 (29.2%) patients dropped out without clear reasons or the group allocation being specified. Selective reporting All studies reported all prespecified outcomes as defined in their methods sections. Other potential sources of bias In the five cross-over trials, no information was provided regarding the randomisation of the order of treatments and no assessment was made of a period effect. Three single dose studies of similar design each conducted the trial over three consecutive nights, without any washout period (Bayley 1976; Hobkirk 1977; Sharma 1978). This may have exposed the trials to the possibility of a carry-over effect. Another common source of bias was failure to report co-interventions (Bayley 1976; Drewes 1998; Vince 1973; Walsh 1996). Effects of interventions See: Summary of findings for the main comparison Muscle relaxant versus control for pain management in rheumatoid arthritis Overall we were able to pool some of the data for three studies of less than one week duration (Bayley 1976; Hobkirk 1977; Sharma 1978) and for three studies of one to six weeks duration (Drewes 1998; Vince 1973; Walsh 1996). Primary outcomes Blinding All studies described the use of a placebo but did not provide specific information about the characteristics of the placebo and in particular whether the placebo was identical to the active treatment. No studies audited participants or study personnel on whether they believed the participants were receiving the active treatment. None of the included studies provided specific information about whether or not study personnel (including outcome assessors) were blinded. This is important and raises the possibility that positive results might be an artefact of physician expectations rather than a true effect. This was illustrated in the Vince 1973 trial where patient global assessment outcomes were the same in both treatment groups however the physician global assessment was higher in the diazepam group than the placebo group. 1) Effectiveness of muscle relaxants - pain intensity No study reported the primary outcome measure of patient reported pain relief of 30% or greater. Available pain data were confined to mean pain VAS or means of ordinal outcomes in all trials. Any muscle relaxant versus placebo When pooled, the short term single dose studies assessing diazepam (52 participants) (Bayley 1976; Hobkirk 1977; Sharma 1978) showed no benefit in mean night pain VAS (0 to 10 cm) scores over the control arm (SMD -0.22, 95% CI to 0.58) (Analysis 1.1, Figure 3). The other three studies of between one and two weeks duration (Drewes 1998; Vince 1973; Walsh 1996) also showed no significant improvement in mean pain over control (SMD -0.20, 95% CI to 0.18) (Analysis 1.2, Figure 4). 14

18 Figure 3. Forest plot of comparison: 1 Muscle relaxant versus control, outcome: 1.1 Pain 24 hrs. Figure 4. Forest plot of comparison: 1 Muscle relaxant versus control, outcome: 1.2 Pain 1-2 weeks. Benzodiazepines versus placebo Three studies assessed the efficacy and safety of a benzodiazepine versus placebo on pain intensity (Bayley 1976; Vince 1973; Walsh 1996) (Analysis 2.1). Two small cross-over trials (Bayley 1976; Vince 1973) compared diazepam with placebo at different time points and reported conflicting results (2 trials, 35 people). Bayley 1976 reported that a single dose of diazepam was superior to placebo in relieving night pain (mean improvement of 0.9 cm, 95% CI to -0.03) on a 10 cm VAS, while Vince 1973 reported no difference in mean pain scores between diazepam and placebo after one week. Two studies compared different benzodiazepines with a placebo over one to two weeks (Vince 1973; Walsh 1996). After two weeks, Vince 1973 again found no significant difference in mean pain scores between diazepam and placebo. Walsh 1996 compared triazolam with placebo in patients with RA and sleep disturbance and also found no significant difference in pain outcomes after two weeks. Pooling these two one week studies using a randomeffects model yielded the same result (SMD -0.19, 95% CI to 0.30) (Analysis 2.1). Non-benzodiazepines versus placebo One small study (1 trial, 41 patients) found no benefit of zopiclone over placebo over two weeks on either present pain intensity (MD -0.20, 95% CI to 0.37) or total pain rating index (MD , 95% CI to 3.05) (Drewes 1998) (Analysis 4.1). Benzodiazepine versus NSAID One short term, single dose cross-over study (18 people) reported no benefit of diazepam over indomethacin in pain outcomes in inpatients with RA (Bayley 1976). However, insufficient data were provided to extract any data regarding this result. Benzodiazepine + NSAID versus NSAID Two small cross-over trials (2 trials, 35 people) evaluated whether or not there was any benefit in the addition of diazepam to an NSAID over taking an NSAID alone (Hobkirk 1977; Sharma 1978). Both trials were small and were three consecutive night trials of inpatients with active disease. Neither trial found an additional benefit of the combination in terms of pain reduction 15

19 compared with NSAID alone. Standard deviations were estimated from a conservative estimate of the P values in each of the trials and when data were pooled the results were the same (SMD -0.14, 95% CI to 1.36) (Analysis 3.1). 2) Safety of muscle relaxants Number of withdrawals due to adverse events There was a paucity of data available (n = 70) from the three studies that compared a benzodiazepine with placebo. Overall there was a trend towards more withdrawals in patients receiving a muscle relaxant but this did not receive statistical significance (RR 2.84, 95% CI 0.31 to 26.08) (Analysis 5.1, Figure 5). No information was provided on withdrawals due to adverse events in Walsh 1996, and there were no events reported in the Bayley 1976 study. Vince 1973 only reported adverse events in 17/24 patients that were entered and completed their trial. It was not specified as to whether the seven patients who did not complete the trial suffered an adverse event. Figure 5. Forest plot of comparison: 5 Muscle relaxant versus control - safety, outcome: 5.1 Withdrawal due to adverse events. There was only one withdrawal due to adverse events in each of the head-to-head trials (over three consecutive nights only) (Hobkirk 1977; Sharma 1978). In the Hobkirk 1977 trial the patient felt generally unwell after one dose of diazepam and indomethacin and in the Sharma 1978 trial one patient withdrew after two nights having developed a plethora of minor symptoms. It was not specified as to which two treatments this patient had received however as two arms of the study included diazepam and different NSAIDs this was included as an experimental event. In the one study of the non-benzodiazepine agent zopiclone versus placebo there were no withdrawals due to adverse events at the end of the two week trial (Drewes 1998). Secondary outcomes Total number of adverse events and serious adverse events (SAEs) Five trials reported adverse event data (Bayley 1976; Drewes 1998; Hobkirk 1977; Sharma 1978; Vince 1973). When pooled there was a trend towards a significant increase in total adverse events only (RR 1.40, 95% CI 0.58 to 3.41) (Analysis 5.2, Figure 6). This was explained by the heterogeneity in the results of the single dose versus longer duration studies. In the single cross-over studies (3 trials, 106 people) evaluating short term (24 hr) outcomes of diazepam versus placebo there was no significant increase in the total number of adverse events (RR 0.78, 95% CI 0.41 to 1.48) (Bayley 1976; Hobkirk 1977; Sharma 1978) (Analysis 5.4). However, In the longer one or two week trials there were significantly more adverse events (RR 4.03, 95% CI 1.08 to 15.10) (Drewes 1998; Vince 1973) (Analysis 5.3, Figure 7). The majority of these were central nervous system effects including drowsiness and dizziness (RR 5.96, 95% CI 1.77 to 20.08) (Analysis 5.5, Figure 8). 16

20 Figure 6. Forest plot of comparison: 5 Muscle relaxant versus control - safety, outcome: 5.2 Total adverse events. Figure 7. Forest plot of comparison: 5 Muscle relaxant versus control - safety, outcome: 5.3 Total adverse events - trials greater than 24hrs duration. 17