BMJ

Confidential: For Review Only Efficacy, acceptability and safety of muscle relaxant medicines for adults with non-specific low back pain: systematic review and meta-analysis

Journal: BMJ

Manuscript ID BMJ-2020-063248.R1

Article Type: Research

BMJ Journal: BMJ

Date Submitted by the 18-Mar-2021 Author:

Complete List of Authors: Cashin, Aidan; University of New South Wales Faculty of Medicine, Prince of Wales Clinical School; Neuroscience Research Australia, Centre for PAIN IMPACT Folly, Thiago; Neuroscience Research Australia, Centre for Pain IMPACT Bagg, Matthew; Neuroscience Research Australia, Centre for Pain IMPACT; University of New South Wales, Prince of Wales Clinical School Wewege, Michael; Neuroscience Research Australia, Centre for Pain IMPACT; University of New South Wales, School of Medical Sciences Jones, Matthew; University of New South Wales Faculty of Medicine, School of Medical Sciences; Neuroscience Research Australia, Centre for Pain IMPACT Ferraro, Michael; Neuroscience Research Australia, Centre for Pain IMPACT; University of New South Wales Faculty of Medicine, School of Medical Sciences Leake, Hayley; University of South Australia, IIMPACT in Health Rizzo, Rodrigo; Neuroscience Research Australia, Centre for Pain IMPACT; University of New South Wales Faculty of Medicine, School of Medical Sciences Schabrun, Siobhan; Neuroscience Research Australia, Centre for Pain IMPACT Gustin, Sylvia; University of New South Wales, School of Psychology, Faculty of Science; Neuroscience Research Australia, Centre for Pain IMPACT Day, Richard; St Vincents Hospital Sydney and UNSW, Clinical Pharmacology & Toxicology; University of New South Wales, St. Vincent’s Clinical School, Faculty of Medicine Williams, Christopher; Hunter New England Population Health; The University of Newcastle, School of Medicine and Public Health McAuley, James; Neuroscience Research Australia, Centre for Pain IMPACT; University of New South Wales Faculty of Medicine, School of Medical Sciences

back pain, analgesics, muscle relaxant medicines, systematic review, Keywords: meta-analysis

https://mc.manuscriptcentral.com/bmj Page 1 of 74 BMJ

1 2 3 4 5 6 7 8 9 10 11 Confidential: For Review Only 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 2 of 74

1 2 3 Efficacy, acceptability and safety of muscle relaxant medicines for adults 4 5 with non-specific low back pain: systematic review and meta-analysis 6 7 Aidan G Cashin (doctoral candidate)1,2, Thiago Folly (research assistant)1, Matthew K Bagg 8 9 (doctoral candidate)1,2,3, Michael A Wewege (doctoral candidate)1,4, Matthew D Jones 10 1,4 1,4 5 11 (lecturer) , Michael C Ferraro (doctoral candidate) , Hayley B Leake (doctoral candidate) , 12 Confidential: For Review Only 1,4 1 13 Rodrigo R N Rizzo (doctoral candidate) , Siobhan M Schabrun (research fellow) , Sylvia M 14 Gustin (associate professor)1,6, Richard Day (professor, medical doctor)7,8, Christopher M 15 16 Williams (associate professor)9,10, James H McAuley (professor)1,4* 17 18 19 Affiliations: 20 21 1. Centre for Pain IMPACT, Neuroscience Research Australia, Sydney, NSW, Australia 22 23 2. Prince of Wales Clinical School, University of New South Wales, Sydney, NSW, Australia 24 25 3. New College Village, University of New South Wales, Sydney, NSW, Australia 26 4. School of Medical Sciences, Faculty of Medicine UNSW, Sydney, NSW, Australia 27 28 5. IIMPACT in Health, University of South Australia, Adelaide, Australia 29 30 6. School of Psychology, Faculty of Science, Sydney, NSW, Australia 31 32 7. Clinical Pharmacology & Toxicology, St. Vincent’s Hospital, Sydney, NSW, Australia 33 8. St. Vincent’s Clinical School, Faculty of Medicine, UNSW, Sydney, Australia 34 35 9. School of Medicine and Public Health, University of Newcastle, Callaghan, Australia 36 37 10. Hunter New England Population Health, Hunter New England Local Health District, 38 Newcastle Australia 39 40 41 42 43 44 *Correspondence to: 45 Prof James H McAuley 46 Neuroscience Research Australia 47 PO Box 1165, Randwick, NSW 2031, Australia 48 Tel: +61 2 9399 1266 49 50 Email: [email protected] 51 52 53 Word Count: 4087 54 55 56 57 58 59 60

Cashin et al. 2021 Page 1 of 31 https://mc.manuscriptcentral.com/bmj Page 3 of 74 BMJ

1 2 3 What is already known on this topic 4 5 6  Muscle relaxant medicines are the third most frequently prescribed medicine for low 7 back pain. 8 9  Clinical practice guidelines provide conflicting recommendations for the use of 10 11 muscle relaxant medicines in the treatment of low back pain. 12 Confidential: For Review Only 13 14 What this study adds 15 16  Very low certainty evidence shows that non-benzodiazepine antispasmodic medicines 17 18 may offer a small, non-clinically important reduction in pain intensity at two weeks or 19 less for acute low back pain. 20 21  Low and very low certainty evidence shows non-benzodiazepine antispasmodic 22 23 medicines may increase the risk of adverse events and may have little to no effect on 24 25 treatment discontinuation compared to control, respectively. 26 27  Large, definitive, placebo-controlled trials are urgently needed to resolve uncertainties 28 for the efficacy and safety of muscle relaxant medicines. 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 2 of 31 https://mc.manuscriptcentral.com/bmj BMJ Page 4 of 74

1 2 3 ABSTRACT 4 5 6 7 Objective 8 To investigate the efficacy, acceptability and safety of muscle relaxant medicines for low 9 10 back pain (LBP). 11 12 Confidential: For Review Only 13 14 Design 15 Systematic review and meta-analysis of randomised controlled trials. 16 17 18 19 Data sources 20 21 MEDLINE, Embase, CINAHL, CENTRAL, ClinicalTrials.gov, clinicialtrialsregister.eu, and 22 WHO ICTRP from inception through to 23rd February 2021. 23 24 25 26 Eligibility criteria for selecting studies 27 Randomised controlled trials of muscle relaxant medicines compared to placebo, usual care, 28 29 waiting list, or no treatment in adults reporting non-specific LBP. 30 31 32 33 Data extraction and synthesis 34 Two reviewers independently identified studies, extracted data, and assessed the risk of bias 35 36 and certainty in the evidence using the Cochrane risk-of-bias tool and GRADE, respectively. 37 38 Random-effects meta-analytic models through restricted maximum likelihood estimation 39 40 were used to estimate pooled effects and 95% confidence intervals. Outcomes included pain 41 intensity (measured on a 0-100 point scale), disability (measured on a 0-100 point scale), 42 43 acceptability (discontinuation of the medicine for any reason during treatment), and safety 44 45 (adverse events, serious adverse events and the number of participants that withdrew from the 46 trial because of an adverse event). 47 48 49 50 Results 51 52 We included 48 trials in the review, of which 31, sampling 6505 participants, were 53 quantitatively analysed. For acute LBP, there is very low certainty evidence that non- 54 55 benzodiazepine antispasmodic medicines reduce pain intensity at ≤ 2 weeks compared to 56 57 control (mean difference [MD] -7.7, 95% confidence interval -12.1 to -3.3) but do not reduce 58 disability at ≤ 2 weeks (MD -3.3, -7.3 to 0.7). Low and very low certainty evidence showed 59 60 non-benzodiazepine antispasmodic medicines may increase the risk of adverse events (risk

Cashin et al. 2021 Page 3 of 31 https://mc.manuscriptcentral.com/bmj Page 5 of 74 BMJ

1 2 3 ratio [RR] 1.6, 1.2 to 2) and may have little to no effect on acceptability (RR 0.8, 0.6 to 1.1) 4 5 compared to control in acute LBP, respectively. The number of trials investigating other 6 7 muscle relaxant medicines and different durations of LBP were small and the certainty in 8 evidence was reduced by a large number of trials at high risk of bias. 9 10 11 12 ConclusionsConfidential: For Review Only 13 14 There is considerable uncertainty in the clinical efficacy and safety of muscle relaxant 15 medicines. Very low and low certainty evidence shows that non-benzodiazepine 16 17 antispasmodic medicines may provide small, but not clinically important reductions in pain 18 19 intensity at or before two weeks and may increase the risk of experiencing an adverse event 20 21 in acute LBP, respectively. Large, high-quality, placebo-controlled trials are urgently needed 22 to resolve uncertainty. 23 24 25 26 Systematic review registration 27 PROSPERO (CRD42019126820); Open Science Framework (https://osf.io/mu2f5/) 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 4 of 31 https://mc.manuscriptcentral.com/bmj BMJ Page 6 of 74

1 2 3 Introduction 4 5 Low back pain (LBP) is a major global public health problem. It has been the leading cause 6 7 of disability worldwide for the past 30 years,[1] placing a burden on individuals, healthcare 8 and society. In the USA, LBP is responsible for the highest total health care spending, 9 10 estimated to be $134.5 billion (95% Confidence Interval [CI], $122.4 to $146.9 billion) in 11 12 2016.[2]Confidential: LBP is a common symptom related-reason For Review to visit a general Only practitioner (GP),[3,4] 13 14 where patients are often prescribed analgesic medicines to manage their pain.[5–7] 15 16 17 Muscle relaxants, a broad class of medicines that include non-benzodiazepine antispasmodic 18 19 and antispastic medicines (table 1), are frequently prescribed in the UK and the USA. In 20 21 England prescriptions exceeded 1.3 million in 2020,[10] and in the USA over 30 million 22 prescriptions of muscle relaxant medicines were recorded for ambulatory care visits in 23 24 2016.[8] Muscle relaxant medicines are the third most frequently prescribed medicine for 25 26 LBP.[5,7–9] There are conflicting recommendations between LBP international clinical 27 practice guidelines outlining the use of muscle relaxant medicines.[10,11] For example, the 28 29 USA guideline recommends non-benzodiazepine antispasmodic medicines as part of first-line 30 31 pharmacological care for acute LBP,[12] the Belgium guideline discourages use,[13] and the 32 33 UK guideline does not make a recommendation.[14] 34 35 36 A systematic review that included five randomised controlled trials (RCTs) (n = 497 37 38 participants) published up to October 2015 provides the most recent evidence that muscle 39 40 relaxant medicines produce a clinically meaningful reduction in pain intensity for people with 41 acute LBP (Mean Difference [MD], -21.3, [95% CI -29.0 to -13.5]).[15] Several large RCTs 42 43 have since been published. Further, this systematic review did not include evidence from 44 45 RCTs in clinical trial registries which may lead to an overestimation of the effect.[16] To 46 address this knowledge gap, we systematically reviewed the evidence to estimate the 47 48 efficacy, acceptability and safety of muscle relaxant medicines compared to placebo, usual 49 50 care or no treatment in adults with LBP. We evaluated the certainty of the evidence 51 52 supporting the findings using the Grading of Recommendations, Assessment, Development 53 and Evaluations (GRADE) approach[17,18] and discuss its clinical relevance. 54 55 56 57 58 Methods 59 60

Cashin et al. 2021 Page 5 of 31 https://mc.manuscriptcentral.com/bmj Page 7 of 74 BMJ

1 2 3 We prospectively registered the protocol on PROSPERO (CRD42019126820) and the Open 4 5 Science Framework (https://osf.io/mu2f5/) on 29 March 2019. We followed the Preferred 6 7 Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) reporting 8 guideline.[19] Protocol deviations are reported in supplemental file 1. 9 10 11 12 Data sourcesConfidential: and searches For Review Only 13 14 We searched MEDLINE, Embase, CINAHL, the Cochrane Central Register of Controlled 15 Trials (CENTRAL), the Cochrane Back and Neck Group’s trial register (through 16 17 CENTRAL), ClinicalTrials.gov, the EU Clinical Trials Register, and the World Health 18 19 Organisation’s (WHO) International Clinical Trial Registry Platform from inception to 23 20 21 February 2021. We developed and piloted the search strategies for bibliographic databases 22 and clinical trial registries using medical subject headings/Emtree and text words for 'LBP', 23 24 'RCTs' and 'spasmolytic muscle relaxant medicines' (supplemental file 2 and 3). We searched 25 26 the reference lists from retrieved full-text articles and previous systematic reviews. We 27 searched PROSPERO for any ongoing or recently completed systematic reviews. 28 29 30 31 Eligibility criteria 32 33 We included records of RCTs that allocated adult participants with non-specific LBP[20] to 34 receive a systemically administered dose of a spasmolytic muscle relaxant medicine[21] 35 36 compared with a non-active control group (i.e., a sham (placebo) medicine, continuation of 37 38 usual care, waiting list, or no treatment). We included RCTs investigating the combination of 39 40 two medicines compared to one medicine alone (e.g., tizanidine and ibuprofen vs ibuprofen 41 alone). We included medicines that were classified pharmacologically as muscle relaxant 42 43 medicines, listed on the WHO Anatomical Therapeutic Chemical (ATC) classification 44 45 system[22] and licensed in either the US (U.S. Food & Drug Administration[23]), Europe 46 (European Medicines Agency[24]) or Australia (Australian Register of Therapeutic 47 48 Goods[25]) as at 29 March 2019 (supplemental file 4). We included trials reported in 49 50 English, Italian, Portuguese, Spanish, German and Dutch. We did not restrict the inclusion of 51 52 trials by the duration of LBP reported, trial publication status, outcomes reported or by the 53 instrument used to assess outcomes. We excluded trials reported in other languages or trials 54 55 that sampled participants with specific spinal pathology (e.g. infection, neoplasm, 56 57 inflammatory disease or fracture)[20] or sciatica.[26] We excluded trials that sampled 58 multiple health conditions unless separate data were available for the participants with non- 59 60 specific LBP.

Cashin et al. 2021 Page 6 of 31 https://mc.manuscriptcentral.com/bmj BMJ Page 8 of 74

1 2 3 4 5 Study selection 6 7 The review team independently screened the titles and abstracts of all identified records in 8 duplicate. We retrieved full-length records of those deemed eligible and screened these again 9 10 to confirm inclusion. The full-length record of a trial registration was defined as the primary 11 12 web-pageConfidential: and all subsidiary pages and files For located Review on the trial registry. Only We resolved 13 14 disagreements through discussion (AGC, MAW, MDJ, MCF, HBL, RRNR) or, when 15 necessary, consultation of a third independent reviewer (MKB or JHM). We contacted 16 17 authors up to three times within a six-week period when further information was required to 18 19 confirm eligibility. 20 21 22 In instances where a trial was linked to multiple record sources, we used an established 23 24 hierarchy preferencing the main published trial report, followed by other published records of 25 26 the trial (e.g. conference abstracts) and lastly, the trial registry record. We classified the trial 27 registry record as the primary record if there was no evidence of publication. 28 29 30 31 Outcomes 32 33 The choice of outcomes was based on the core outcome domains for clinical trials in 34 LBP,[27] and those of other reviews of analgesic medicines for LBP.[28–30] The primary 35 36 outcomes were pain intensity and acceptability (satisfaction with the treatment regime 37 38 measured by the number of patients who discontinued the treatment for any reason). 39 40 Disability, adverse events (as defined by each study), serious adverse events (as defined by 41 each study) and withdrawal from treatment because of adverse effects (tolerability) were 42 43 secondary outcomes. 44 45 46 Data extraction 47 48 Using a standardised, piloted form, two reviewers independently extracted data on the trial 49 50 characteristics, participants, interventions, comparisons and outcomes from each trial. In the 51 52 absence of data, we transformed or estimated measures of variance using the 53 recommendations from section 6.5.2 in the Cochrane Handbook for Systematic Reviews of 54 55 Interventions.[31] Briefly, we transformed standard errors or 95% confidence intervals for 56 57 group-level estimates to standard deviations using equations from section 6.5.2.2.[31] If 58 studies reported mean difference between groups and a p-value, we calculated the between- 59 60 group standard error using equations from section 6.5.2.3 and used the between-group

Cashin et al. 2021 Page 7 of 31 https://mc.manuscriptcentral.com/bmj Page 9 of 74 BMJ

1 2 3 estimates in the meta-analysis.[31] When no measure of variance was reported, we imputed a 4 5 conservative standard deviation of 30. We resolved disagreements for data extraction through 6 7 discussion, or with arbitration by a third reviewer if necessary. We contacted authors up to 8 three times over a period of six weeks to request data not reported in the RCT. 9 10 11 12 Risk of biasConfidential: and certainty of evidence For Review Only 13 14 Two independent reviewers appraised study-level risk of bias using the Cochrane ‘risk of 15 bias’ tool [32] and recommendations from Furlan et al. 2015.[33] We assessed 13 criteria 16 17 across the risk of bias domains selection, performance, attrition, detection, reporting and 18 19 other sources of bias.[33] We used the criteria in Furukawa et al. 2016.[34] to determine 20 21 overall study risk of bias for each trial because recommendations are not available from 22 Furlan et al.[33] Two reviewers independently formed judgements of certainty in the 23 24 evidence for each analysis using the Grading of Recommendations, Assessment, 25 26 Development and Evaluation (GRADE) system.[17,18] The certainty in the evidence was 27 initially classified as ‘high’ and was downgraded according to the following recommended 28 29 criteria: risk of bias, inconsistency, imprecision and publication bias (supplemental file 5). 30 31 The certainty in the evidence was defined as ‘high’, ‘moderate’, ‘low’ or ‘very low’ certainty. 32 33 Disagreements between appraisals of risk of bias and certainty of evidence were resolved 34 through discussion or when required, with arbitration by a third reviewer. 35 36 37 38 Data synthesis and analysis 39 40 We conducted meta-analyses of trials for each outcome using the available data for 41 immediate term (≤ 2 weeks post randomisation) and short-term follow-up (3-13 weeks post 42 43 randomisation). We chose the time closest to six weeks when data for multiple time points 44 45 were available for the short-term follow-up. We stratified all analyses by the clinical utility of 46 the muscle relaxant medicine (antispastics, non-benzodiazepine antispasmodics, 47 48 benzodiazepines and miscellaneous) and the duration of LBP observed in the included trials; 49 50 acute (0-6 weeks), sub-acute (6-12 weeks), chronic (>12 weeks), and mixed (participants 51 52 with multiple symptom durations). We followed guidance[31] to incorporate trials with 53 multiple comparisons; dividing the control group sample size by the number of trial arms. As 54 55 benchmarks for clinically important effects are usually expressed on a 0-100 scale[35,36] and 56 57 to facilitate clinical interpretation of results,[37] we converted aggregate outcome data 58 (measure of central tendency and dispersion), for pain and disability, to a common 0 (no 59 60 pain/disability) to 100 (worst pain/disability) scale. Details on the pain and disability

Cashin et al. 2021 Page 8 of 31 https://mc.manuscriptcentral.com/bmj BMJ Page 10 of 74

1 2 3 measures used by studies, and conversion procedures are provided in supplemental files 6 4 5 and 7. We considered a difference in favour of muscle relaxant medicines of at least 10 points 6 7 for pain and disability to be the minimal clinically important effect.[37,38] This threshold has 8 been used in other reviews of analgesic medicines for LBP.[28,30,39] 9 10 11 12 We fit random-effectsConfidential: meta-analytic models For through Review Restricted Maximum Only Likelihood 13 14 Estimation, using dmetar in R (version 3.6.1).[40] We expressed effects for continuous 15 outcomes with the mean between-group difference and accompanying 95% confidence 16 17 intervals and effects for binary outcomes using the risk ratio and accompanying 95% 18 19 confidence intervals. We estimated the Q statistic and the between-study variance (τ2) from 20 21 each analysis and used these values to calculate 95% prediction intervals for the pooled effect 22 and I² values. We used these measures to form judgements about heterogeneity in 23 24 conjunction with visual inspection of the distribution of effect sizes in the forest plots. We 25 26 formed judgements about publication bias for each meta-analysis by visually inspecting 27 funnel plots and considering the proportion of trials included from trial registry records for 28 29 that outcome. 30 31 32 33 Planned investigation of heterogeneity 34 We conducted a planned subgroup analysis to explore whether heterogeneity varied by 35 36 prescribed dose. Dose had three levels: standard dose, above standard dose or below standard 37 38 dose, according to the Prescribers Digital Reference,[41] Monthly Index of Medical 39 40 Specialities[42] or Australian Medicines Handbook.[43] 41 42 43 Sensitivity analyses 44 45 We conducted sensitivity analyses to assess the influence on effect estimates of trials with 46 unclear definitions of non-specific LBP, trials where measures of variance were imputed, 47 48 trials at high risk of bias, trials reported as trial registry records, trials without a placebo 49 50 comparison, and trials investigating the muscle relaxant medicine carisoprodol. This was 51 52 done by repeating the main analyses without the relevant trials included. 53 54 55 Patient and public involvement 56 57 There was no direct patient and public involvement in this study due to a lack of funding, 58 although we did speak to patients about the study and we asked a member of the public to 59 60

Cashin et al. 2021 Page 9 of 31 https://mc.manuscriptcentral.com/bmj Page 11 of 74 BMJ

1 2 3 read our manuscript after submission. We plan to disseminate the results of this review to the 4 5 relevant patient organisations. 6 7 8 9 10 Results 11 12 We identifiedConfidential: 3362 records, removed 215 For duplicates Review and screened 3147 Only records during title 13 14 and abstract screening. We included 48 trials, 34 were peer-reviewed journal articles,[44–79] 15 two were conference abstracts[55,71] and 12 were trial registry records[80–91] (fig 1 and 16 17 supplemental file 8). Thirty-one trials (two trial registry records and one conference abstract), 18 19 including a total 6505 participants, contributed data to the meta-analyses. Trials that did not 20 21 provide eligible data for meta-analyses of the primary outcome pain intensity are described in 22 supplemental file 9. 23 24 25 26 27 Study characteristics 28 The included trials investigated 18 different muscle relaxant medicines, most commonly non- 29 30 benzodiazepine antispasmodics (n = 29), miscellaneous (n =11), antispastics (n = 5), and 31 32 benzodiazepines (n = 3). The muscle relaxant medicines investigated were administered 33 orally in 35 trials, by intramuscular injections in 10 trials, and by intravenous injection in one 34 35 trial. Two trials used mixed administration of a starting intramuscular injection and 36 37 subsequent oral doses. Most trials (n = 31) compared a muscle relaxant medicine to placebo. 38 39 40 Thirty-four trials sampled participants with acute LBP, two with sub-acute LBP, and eight 41 42 with chronic LBP. Two trials investigated participants with both acute and sub-acute LBP 43 44 (mixed duration sample) and two trials did not report duration of LBP. 45 46 47 Risk of bias 48 49 We assessed completed trials (n = 38) for overall risk of bias; eight trials were assessed at 50 51 low risk of bias, three at moderate risk of bias, and 27 at high risk of bias. The most common 52 reasons for high risk of bias were from attrition bias (failure to report intention to treat 53 54 effects, n=8), performance bias (inadequate blinding of participants, n = 6 and care providers, 55 56 n = 7) and detection bias (inadequate blinding of outcome assessors, n = 6). There was also 57 58 unclear risk of selection bias from inadequate reporting of allocation concealment (n = 32), 59 and random sequence generation (n = 24) (supplemental file 10). 60

Cashin et al. 2021 Page 10 of 31 https://mc.manuscriptcentral.com/bmj BMJ Page 12 of 74

1 2 3 4 5 Efficacy and acceptability 6 7 Acute LBP 8 Seventeen trials (24 comparisons) determined the efficacy of muscle relaxant medicines at 9 10 immediate term (≤ 2 weeks) and four trials (7 comparisons) for short-term (3-13 weeks) 11 12 follow-upConfidential: for acute LBP. Thirteen trials (16For comparisons) Review determined Only the acceptability of 13 14 muscle relaxant medicines for acute LBP. Non-benzodiazepine antispasmodic medicines 15 reduced pain intensity at ≤ 2 weeks compared to control (MD -7.7, -12.1 to -3.3; 16 trials, 16 17 4546 participants; very low certainty evidence) (fig 2 and table 2). Non-benzodiazepine 18 19 antispasmodic medicines did not reduce pain intensity compared to control at 3-13 weeks 20 21 (MD 0.6, -4.5 to 5.7; 3 trials, 612 participants; moderate certainty evidence) or disability at ≤ 22 2 weeks (MD -3.3, -7.3 to 0.7; 7 trials, 2438 participants; very low certainty evidence) and 3- 23 24 13 weeks (MD 4.3, -1.4 to 10.1; 2 trials, 422 participants; moderate certainty evidence) (table 25 26 2-3 and supplemental files 11-13). There was no difference in the acceptability of non- 27 benzodiazepine antispasmodic medicines compared to control (risk ratio [RR] 0.8, 0.6 to 1.1; 28 29 13 trials, 2834 participants; very low certainty evidence) (table 4, supplemental file 14). 30 31 Evidence ranging from moderate to low certainty showed no benefit of antispastic and 32 33 benzodiazepine medicines compared to control for pain and disability at immediate (≤ 2 34 weeks) and short-term (3-13 weeks) follow-up, except for reduction in disability for 35 36 benzodiazepine medicines at 3-13 weeks (MD -6.9, -12.1 to -1.7; 1 trial, 103 participants; 37 38 moderate certainty evidence) (table 2-3, supplemental files 11-13). 39 40 41 Subacute LBP 42 43 One trial determined the efficacy of muscle relaxant medicines on pain intensity at short-term 44 45 follow-up for subacute LBP. Miscellaneous muscle relaxant medicines (botulinum toxin) did 46 not reduce pain intensity compared to control (MD -19, -41.9 to 3.9; one trial, 28 participants; 47 48 very low certainty evidence) (table 2, supplemental file 11). 49 50 51 52 Chronic LBP 53 Two trials (3 comparisons) determined the efficacy of muscle relaxant medicines on pain 54 55 intensity at short-term follow-up for chronic LBP. Three trials (4 comparisons) determined 56 57 the acceptability of muscle relaxant medicines for chronic LBP. Evidence ranging from 58 moderate to very low certainty showed no benefit compared to control for antispastic and 59 60 miscellaneous muscle relaxant medicines for pain, disability and acceptability except for pain

Cashin et al. 2021 Page 11 of 31 https://mc.manuscriptcentral.com/bmj Page 13 of 74 BMJ

1 2 3 intensity at 3-13 weeks from miscellaneous muscle relaxant medicines (eszoplicone) (MD - 4 5 19.9, -31.5 to -8.3; 1 trial, 52 participants; moderate certainty evidence) (table 2-4, 6 7 supplemental file 11, 13-14). 8 9 10 Mixed LBP 11 12 Two trialsConfidential: (2 comparisons) determined theFor efficacy Review of muscle relaxant Only medicines at 13 14 immediate term and one trial for short-term follow-up for mixed LBP. Non-benzodiazepine 15 antispasmodic medicines reduced pain intensity at ≤ 2 weeks compared to control (MD -4.4, - 16 17 6.9 to -1.9; 2 trials, 617 participants; low certainty) but not at 3-13 weeks (MD -5.8, -13.8 to 18 19 2.2; 1 trial, 329 participants; very low certainty evidence) (fig 2, table 2 and supplemental file 20 21 11). Non-benzodiazepine antispasmodic medicines also reduced disability at ≤ 2 weeks 22 compared to control (MD -19.2, -27.7 to -10.7; 1 trial, 329 participants; low certainty 23 24 evidence) (table 3 and supplemental file 12). 25 26 27 Safety 28 29 Acute LBP 30 31 Twenty-two trials (28 comparisons) determined the safety of muscle relaxant medicines for 32 33 acute LBP. The type and reporting of adverse and serious adverse events varied across trials. 34 Compared to control, non-benzodiazepine antispasmodic medicines increased the risk of an 35 36 adverse event (RR 1.6, 1.2 to 2; 16 trials, 3404 participants; low certainty evidence), but not 37 38 serious adverse events (RR 2.3, 0.3 to 20.8; 2 trials, 830 participants; very low certainty 39 40 evidence) (table 5, supplemental file 15-16). Antispastic medicines also increased the risk of 41 an adverse event (RR 2, 1.1 to 3.8; 2 trials, 290 participants; moderate certainty evidence), 42 43 but benzodiazepine medicines did not (RR 1.8, 0.9 to 3.6; 2 trials, 159 participants; low 44 45 certainty evidence) (table 5, supplemental file 15). Participants receiving antispastic 46 medicines were more likely to discontinue treatment due to an adverse event (RR 34.6, 2.1 to 47 48 568; 1 trial, 195 participants; very low certainty), whereas participants receiving non- 49 50 benzodiazepine antispasmodic medicines were not (RR 1.5, 0.6 to 3.5; 5 trials, 1641 51 52 participants; very low certainty) (table 5, supplemental file 17). 53 54 55 Chronic LBP 56 57 Two trials (2 comparisons) determined the safety of muscle relaxant medicines for chronic 58 LBP. Compared to control, there was no difference in the risk of experiencing an adverse 59 60

Cashin et al. 2021 Page 12 of 31 https://mc.manuscriptcentral.com/bmj BMJ Page 14 of 74

1 2 3 event with miscellaneous muscle relaxant medicines (RR 1.5, 0.4 to 5.7; 2 trials, 95 4 5 participants; moderate certainty evidence) (table 5, supplemental file 15). 6 7 8 Mixed LBP 9 10 One trial determined the safety of muscle relaxant medicines for mixed LBP. Compared to 11 12 control, Confidential:there was no difference in the risk For of experiencing Review an adverse Only event with non- 13 14 benzodiazepine antispasmodic medicines (RR 1.6, 0.6 to 4.3; 1 trial, 329 participants; very 15 low certainty evidence) (table 5, supplemental file 15). 16 17 18 19 Subgroup and sensitivity analyses 20 21 Detailed results for the subgroup and sensitivity analyses are presented in supplemental files 22 18 and 19, respectively. Due to a lack of data, sensitivity analyses were only conducted for 23 24 non-benzodiazepine antispasmodic muscle relaxant medicines for acute LBP. The sensitivity 25 26 analyses did not explain heterogeneity in the disability (≤ 2 weeks), acceptability, adverse 27 events and tolerability outcomes. Excluding trials from clinical trial registries (MD -10.2, - 28 29 15.6 to -4.7) or trials without a placebo comparator (MD -11, -17 to -5.1) changed the 30 31 estimated effect for pain intensity (≤ 2 weeks) to within the minimum clinically important 32 33 difference. However, excluding trials at high risk of bias reduced the effect to zero (MD 0.2, - 34 4.9 to 5.4). 35 36 37 38 39 40 Discussion 41 We found very low certainty evidence that non-benzodiazepine antispasmodic medicines 42 43 may reduce pain intensity at two weeks or less for acute LBP. This effect is small, <8 points 44 45 on a 0–100 point scale, and does not meet common thresholds to be clinically meaningful. 46 Non-benzodiazepine antispasmodic medicines may have little to no effect on pain intensity at 47 48 3-13 weeks, and on disability, however the certainty of evidence ranged from moderate to 49 50 very low. No trials evaluated the effect of muscle relaxant medicines on long term outcomes. 51 52 Low and very low certainty evidence showed non-benzodiazepine antispasmodic medicines 53 may increase the risk of adverse events and may have little to no effect on treatment 54 55 discontinuation, respectively. The number of trials investigating other muscle relaxant 56 57 medicines was small. The certainty in evidence was reduced by a large number of trials at 58 high risk of bias. 59 60

Cashin et al. 2021 Page 13 of 31 https://mc.manuscriptcentral.com/bmj Page 15 of 74 BMJ

1 2 3 Strengths and weaknesses of the study 4 5 This systematic review was prospectively registered and reported in line with PRISMA.[19] 6 7 We included a broad scope of licensed muscle relaxant medicines evaluated in RCTs as they 8 provide the best evidence on the efficacy and safety of currently used muscle relaxant 9 10 medicines in clinical practice. We included findings from 48 trials and provide the most up- 11 12 to-date reviewConfidential: of muscle relaxant medicines For in LBP. Review We used the Cochrane Only risk of bias tool 13 14 and published recommendations from the Cochrane Back and Neck Group[33] to assess 15 study level risk of bias and evaluated the certainty in the evidence using the GRADE 16 17 system.[18] Finally, we searched clinical trial registries for relevant trials, whereas previous 18 19 reviews did not.[15,92] 20 21 22 Inadequate reporting and a failure to respond to data requests meant some relevant trials were 23 24 not included in each meta-analysis. We restricted the inclusion of studies based on 25 26 publication language. Although we included trials published in English, Italian, Portuguese, 27 Spanish, German and Dutch, we acknowledge that we may have missed some relevant trials. 28 29 We relied on the definition of adverse and serious adverse events as reported from the 30 31 included trials, however definitions may have varied between trials. We included trials in 32 33 which participants received co-administration of additional analgesic medicines. Although 34 most included trials were compared to placebo, the inclusion of other analgesic medicines 35 36 may influence the interpretation of findings. Finally, interpretation of the outcome 37 38 acceptability has limitations because participants could discontinue treatment for any reason 39

40 including recovery. 41 42 43 Evidence update 44 45 Our review updates the evidence for muscle relaxant medicines for adults with LBP. For 46 example, we included 31 trials (6505 participants) in the quantitative analysis, whereas the 47 48 most recent systematic review included 15 trials (3362 participants).[15] Although the 49 50 previous systematic reviews,[15,92] and three clinical practice guidelines[11] endorse non- 51 52 benzodiazepine antispasmodic medicines as an effective treatment for acute LBP, our 53 findings show that there is considerable uncertainty in their clinical effectiveness and safety. 54 55 56 57 For this review we pooled data for all non-benzodiazepine antispasmodic medicines 58 according to the shared clinical utility in managing acute LBP.[21] The choice of muscle 59 60 relaxant medicine and frequency of prescription by a GP varies between countries.[9,93] For

Cashin et al. 2021 Page 14 of 31 https://mc.manuscriptcentral.com/bmj BMJ Page 16 of 74

1 2 3 example carisoprodol, commonly prescribed in the USA,[8] is no longer marketed in several 4 5 Europeans countries[94] and in the UK[95] due to an increased risk of abuse and addiction. 6 7 We repeated the primary analyses with trials that investigated carisoprodol removed and 8 showed that effects were comparable. 9 10 11 12 MeaningConfidential: of the study For Review Only 13 14 Although the observed effect of non-benzodiazepine antispasmodic medicines in reducing 15 pain compared to control at less than two weeks was statistically significant, the magnitude of 16 17 the effect was too small to be considered clinically important. However, the upper limits of 18 19 the confidence interval do not exclude a clinically meaningful effect on pain intensity. The 20 21 modest overall effect is reported at group level, which could still mean that some individuals 22 gain a worthwhile benefit and some do not.[96] 23 24 25 26 We identified important heterogeneity in the effect of non-benzodiazepine antispasmodic 27 medicines on pain with the prediction interval spanning -26.5 to 11.1. Heterogeneity was 28 29 likely increased by the inclusion of trials reporting unusually large effects such as that of 30 31 Hindle et al., 1972[57] (-48.5, -70.7 to -26.3), where the placebo comparison had no change 32 33 from baseline which is atypical of the natural history of recovery for acute LBP.[97] The 34 subgroup and sensitivity analyses were unable to explain the heterogeneity observed in the 35 36 pooled effect for pain. Restricting the analysis to compare non-benzodiazepine antispasmodic 37 38 medicines with placebo demonstrated a statistically significant effect greater than threshold 39 40 for the minimally clinical important difference. However, we advise caution when 41 interpreting these findings as the removal of high risk of bias trials reduced the effect to zero. 42 43 44 45 Implications for clinical practice and policy 46 International clinical practice guidelines provide conflicting recommendations outlining the 47 48 use of muscle relaxant medicines [10]; 6/15 (40%) clinical practice guidelines recommend 49 50 muscle relaxant medicines to manage LBP, 5/15 (33%) do not and 4/15 (27%) do not offer a 51 52 recommendation.[11] Our review shows that there is uncertainty in the efficacy and safety of 53 muscle relaxant medicines. Although non-benzodiazepine antispasmodics may reduce pain 54 55 intensity at two weeks or less for acute LBP, the effect is unlikely to be considered clinically 56 57 important. In addition, non-benzodiazepine antispasmodics may increase the risk of 58 experiencing an adverse event (commonly dizziness, drowsiness, headache and nausea[98]). 59 60 However, the low to very low certainty of evidence does not allow any firm

Cashin et al. 2021 Page 15 of 31 https://mc.manuscriptcentral.com/bmj Page 17 of 74 BMJ

1 2 3 recommendations. We would encourage clinicians to share this information about the 4 5 efficacy and safety of muscle relaxant medicines with patients to allow them to make 6 7 informed treatment decisions. 8 9 10 Unanswered questions and future research 11 12 Large, definitive,Confidential: placebo-controlled trials For are urgently Review needed to evaluate Only the efficacy and 13 14 safety of muscle relaxant medicines. New trials should follow the core outcome set for non- 15 specific LBP,[99] and recommendations of the Initiative on Methods, Measurement, and Pain 16 17 Assessment in Clinical Trials (IMMPACT),[100] to improve pooling results and 18 19 comparability between trials. Future trials should also endeavour to adhere to methodological 20 21 safeguards to reduce bias and transparently report results following CONSORT.[101] 22 23 24 Although muscle relaxant medicines are typically prescribed for short-term use, the effects of 25 26 long-term use of muscle relaxant medicines are not known. High quality data is required to 27 evaluate the efficacy and safety of long-term use. This is particularly important considering 28 29 the risk of dependency and abuse associated with muscle relaxant medicines has been 30 31 observed from indirect evidence.[92,98] 32 33 34 Conclusions 35 36 This systematic review found very low certainty evidence that non-benzodiazepine 37 38 antispasmodic medicines for the treatment of acute LBP may provide a small and not 39 40 clinically meaningful improvement in pain intensity at two weeks or less. The risk of adverse 41 events but not serious adverse events may be increased with non-benzodiazepine 42 43 antispasmodic medicines, although the evidence ranges from low to very low certainty. 44 45 Large, high-quality, placebo-controlled trials are urgently needed to resolve uncertainties 46 about the efficacy and safety of muscle relaxant medicines for LBP. 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 16 of 31 https://mc.manuscriptcentral.com/bmj BMJ Page 18 of 74

1 2 3 Contributors 4 5 TF, MKB and JHM conceived the idea for the project. TF, MKB, MAW, MCF, SMS, SMG, 6 7 RD and JHM contributed to the project design and protocol development. TF conducted the 8 search. AGC, TF, MKB, MDJ, MAW, MCF, RRNR, HBL conducted the study selection, 9 10 data extraction and quality appraisal. AGC and MAW analysed the data. AGC, MAW and 11 12 JHM hadConfidential: full access to all the data in the For study and Review takes responsibility Only for the integrity of the 13 14 data and the accuracy of the data analysis. AGC wrote the first draft of the manuscript. All 15 authors provided substantive feedback on the manuscript and have read and approved the 16 17 final version. The corresponding author (the manuscript’s guarantor) attests that all listed 18 19 authors meet authorship criteria and that no others meeting the criteria have been omitted. 20 21 22 Funding 23 24 This research received no specific grant from any funding agency in the public, commercial, 25 26 or not-for-profit sectors. AGC is supported by the University of New South Wales Prince of 27 Wales Clinical School Postgraduate Research Scholarship and a NeuRA PhD Candidature 28 29 Supplementary Scholarship. MKB is supported by a NeuRA PhD Candidature Scholarship 30 31 and Supplementary Scholarship and was additionally funded during this work by an 32 33 Australian Government Research Training Program Scholarship and a UNSW Research 34 Excellence Award. MAW is supported by a University Postgraduate Award and School of 35 36 Medical Sciences Top-Up Scholarship from the University of New South Wales, and a 37 38 Postgraduate Scholarship from the National Health and Medical Research Council of 39 40 Australia. MCF is supported by an Australian Medical Research Future Fund Grant 41 GNTID1170205. HBL is supported by Australian Government post-graduate award. RRNR 42 43 is supported by the University of New South Wales School of Medical Sciences Postgraduate 44 45 Research Scholarship and a NeuRA PhD Candidature Supplementary Scholarship. SMS 46 receives salary support from the National Health and Medical Research Council of Australia 47 48 (#1105040). SMG is supported by the Rebecca L. Cooper Medical Research Foundation. 49 50 JHM receives project funding support from the National Health and Medical Research 51 52 Council and the Medical Research Future Fund of Australia. 53 54 55 Competing interests 56 57 All authors have completed the ICMJE uniform disclosure form 58 at www.icmje.org/coi_disclosure.pdf and declare: no support from any organisation for the 59 60 submitted work; no financial relationships with any organisations that might have an interest

Cashin et al. 2021 Page 17 of 31 https://mc.manuscriptcentral.com/bmj Page 19 of 74 BMJ

1 2 3 in the submitted work in the previous three years; no other relationships or activities that 4 5 could appear to have influenced the submitted work. 6 7 8 Ethical approval 9 10 Not required. 11 12 Confidential: For Review Only 13 14 Data sharing 15 The dataset used and analysed during this study, and the accompanying code is available 16 17 from the corresponding author upon reasonable request. 18 19 20 21 Transparency 22 All authors affirm that the manuscript is an honest, accurate, and transparent account of the 23 24 study being reported; no important aspects of the study have been omitted; and any 25 26 discrepancies from the study as originally planned (and, if relevant, registered) have been 27 explained. 28 29 30 31 Dissemination to participants and related patient and public communities 32 33 We will disseminate our findings to patient organisations and traditional media and social 34 media outlets. 35 36 37 38 Patient consent for publication 39 40 Not required. 41 42 43 Copyright statement/licence for publication 44 45 This is an Open Access article distributed in accordance with the Creative Commons 46 Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, 47 48 remix, adapt, build upon this work non-commercially, and license their derivative works on 49 50 different terms, provided the original work is properly cited and the use is non- commercial. 51 52 See: http://creativecommons.org/licenses/by-nc/4.0/. 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 18 of 31 https://mc.manuscriptcentral.com/bmj BMJ Page 20 of 74

1 2 3 References 4 5 1 James SL, Abate D, Abate KH, et al. Global, regional, and national incidence, 6 prevalence, and years lived with disability for 354 diseases and injuries for 195 7 countries and territories, 1990–2017: a systematic analysis for the Global Burden of 8 Disease Study 2017. Lancet 2018;392:1789–858. doi:10.1016/S0140-6736(18)32279- 9 10 7 11 2 Dieleman JL, Cao J, Chapin A, et al. US Health Care Spending by Payer and Health 12 Condition,Confidential: 1996-2016. JAMA 2020; For323:863–84. Review doi:10.1001/jama.2020.0734 Only 13 3 Deyo RA, Weinstein JN. Low Back Pain. N Engl J Med 2001;344:363–70. 14 doi:10.1056/NEJM200102013440508 15 4 Bernstein IA, Malik Q, Carville S, et al. Low back pain and sciatica: Summary of 16 17 NICE guidance. BMJ 2017;356. doi:10.1136/bmj.i6748 18 5 Gore M, Tai KS, Sadosky A, et al. Use and Costs of Prescription Medications and 19 Alternative Treatments in Patients with Osteoarthritis and Chronic Low Back Pain in 20 Community-Based Settings. Pain Pract 2012;12:550–60. doi:10.1111/j.1533- 21 2500.2012.00532.x 22 6 Hart OR, Uden RM, McMullan JE, et al. A study of National Health Service 23 management of chronic osteoarthritis and low back pain. Prim Health Care Res Dev 24 25 2015;16:157–66. doi:10.1017/S1463423614000140 26 7 Ivanova JI, Birnbaum HG, Schiller M, et al. Real-world practice patterns, health-care 27 utilization, and costs in patients with low back pain: The long road to guideline- 28 concordant care. Spine J 2011;11:622–32. doi:10.1016/j.spinee.2011.03.017 29 8 Soprano SE, Hennessy S, Bilker WB, et al. Assessment of Physician Prescribing of 30 Muscle Relaxants in the United States, 2005-2016. JAMA Netw open 2020;3:e207664. 31 doi:10.1001/jamanetworkopen.2020.7664 32 33 9 Mafi JN, McCarthy EP, Davis RB, et al. Worsening trends in the management and 34 treatment of back pain. JAMA Intern Med 2013;173:1573–81. 35 doi:10.1001/jamainternmed.2013.8992 36 10 Schreijenberg M, Koes BW, Lin CWC. Guideline recommendations on the 37 pharmacological management of non-specific low back pain in primary care–is there a 38 need to change? Expert Rev Clin Pharmacol 2019;12:145–57. 39 40 doi:10.1080/17512433.2019.1565992 41 11 Oliveira CB, Maher CG, Pinto RZ, et al. Clinical practice guidelines for the 42 management of non-specific low back pain in primary care: an updated overview. Eur 43 Spine J 2018;27:2791–803. doi:10.1007/s00586-018-5673-2 44 12 Qaseem A, Wilt TJ, McLean RM, et al. Noninvasive Treatments for Acute, Subacute, 45 and Chronic Low Back Pain: A Clinical Practice Guideline From the American 46 College of Physicians. Ann Intern Med 2017;166:514. doi:10.7326/M16-2367 47 48 13 Van Wambeke Leuven PU, President G, Desomer A, et al. Low Back Pain and 49 radicular pain: assessment and management. Belgium: Belgian Health Care 50 Knowledge Centre 2017. 51 14 National Institute for Health and Care Excellence. Low Back Pain and Sciatica in 52 Over 16s: Assessment and Management. London, UK: National Institute for Health 53 and Care Excellence 2016. 54 https://www.nice.org.uk/guidance/ng59/chapter/Recommendations#non-invasive- 55 56 treatments-for-low-back-pain-and-sciatica 57 15 Abdel Shaheed C, Maher CG, Williams KA, et al. Efficacy and tolerability of muscle 58 relaxants for low back pain: Systematic review and meta-analysis. Eur J Pain 59 2017;21:228–37. doi:10.1002/ejp.907 60 16 Bagg MK, O’Hagan E, Zahara P, et al. Reviews may overestimate the effectiveness of

Cashin et al. 2021 Page 19 of 31 https://mc.manuscriptcentral.com/bmj Page 21 of 74 BMJ

1 2 3 medicines for back pain: systematic review and meta-analysis. J Clin Epidemiol 4 Published Online First: 6 December 2019. doi:10.1016/j.jclinepi.2019.12.006 5 6 17 Balshem H, Helfand M, Schünemann HJ, et al. GRADE guidelines: 3. Rating the 7 quality of evidence. J Clin Epidemiol 2011;64:401–6. 8 doi:10.1016/j.jclinepi.2010.07.015 9 18 Guyatt GH, Oxman AD, Vist GE, et al. GRADE: an emerging consensus on rating 10 quality of evidence and strength of recommendations. BMJ 2008;336:924–6. 11 doi:10.1136/bmj.39489.470347.AD 12 Confidential: For Review Only 19 Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting items for systematic reviews 13 14 and meta-analyses: the PRISMA statement. BMJ 2009;339:b2535–b2535. 15 doi:10.1136/bmj.b2535 16 20 Maher C, Underwood M, Buchbinder R. Non-specific low back pain. Lancet 17 2017;389:736–47. doi:10.1016/S0140-6736(16)30970-9 18 21 Trevor A, G KB, Knuidering-Hall M. Skeletal Muscle Relaxants. In: Katzung & 19 Trevor’s Pharmacology: Examination & Board Review. 2018. 20 21 22 WHO Collaborating Centre for Drug Statistics Methodology. ATC/DDD Index. 22 2020.https://www.whocc.no/atc_ddd_index/ 23 23 U.S. Food and Drug Administration. Drugs@FDA. 24 2020.https://www.fda.gov/drugs/development-approval-process-drugs/drug-approvals- 25 and-databases 26 24 European Medicines Agency. Medicines. 27 2020.https://www.ema.europa.eu/en/medicines 28 29 25 Australian Government Department of Health Therapeutic Goods Administration. 30 Australian Register of Therapeutic Goods. 2020.https://www.tga.gov.au/australian- 31 register-therapeutic-goods 32 26 Koes BW, van Tulder MW, Peul WC. Diagnosis and treatment of sciatica. BMJ 33 2007;334:1313–7. doi:10.1136/bmj.39223.428495.BE 34 27 Chiarotto A, Deyo RA, Terwee CB, et al. Core outcome domains for clinical trials in 35 non-specific low back pain. Eur Spine J 2015;24:1127–42. doi:10.1007/s00586-015- 36 37 3892-3 38 28 Machado GC, Maher CG, Ferreira PH, et al. Non-steroidal anti-inflammatory drugs 39 for spinal pain: A systematic review and meta-analysis. Ann Rheum Dis 40 2017;76:1269–78. doi:10.1136/annrheumdis-2016-210597 41 29 Abdel Shaheed C, Maher CG, Williams KA, et al. Efficacy, Tolerability, and Dose- 42 Dependent Effects of Opioid Analgesics for Low Back Pain. JAMA Intern Med 43 44 2016;176:958. doi:10.1001/jamainternmed.2016.1251 45 30 Ferreira GE, McLachlan AJ, Lin CWC, et al. Efficacy and safety of antidepressants for 46 the treatment of back pain and osteoarthritis: Systematic review and meta-analysis. 47 BMJ 2021;372. doi:10.1136/bmj.m4825 48 31 Higgins JP, Thomas J, Chandler J, et al. Cochrane Handbook for Systematic Reviews 49 of Interventions. 2nd Editio. Chichester (UK): John Wiley & Sons 2019. 50 https://training.cochrane.org/handbook/current (accessed 17 Feb 2020). 51 52 32 Higgins JPT, Altman DG, Gøtzsche PC, et al. The Cochrane Collaboration’s tool for 53 assessing risk of bias in randomised trials. BMJ 2011;343. doi:10.1136/bmj.d5928 54 33 Furlan AD, Malmivaara A, Chou R, et al. 2015 updated method guideline for 55 systematic reviews in the Cochrane Back and Neck Group. Spine 2015;40:1660–73. 56 doi:10.1097/BRS.0000000000001061 57 34 Furukawa TA, Salanti G, Atkinson LZ, et al. Comparative efficacy and acceptability 58 of first-generation and second-generation antidepressants in the acute treatment of 59 60 major depression: Protocol for a network meta-analysis. BMJ Open 2016;6.

Cashin et al. 2021 Page 20 of 31 https://mc.manuscriptcentral.com/bmj BMJ Page 22 of 74

1 2 3 doi:10.1136/bmjopen-2015-010919 4 35 Dworkin RH, Turk DC, Wyrwich KW, et al. Interpreting the Clinical Importance of 5 6 Treatment Outcomes in Chronic Pain Clinical Trials: IMMPACT Recommendations. 7 J. Pain. 2008;9:105–21. doi:10.1016/j.jpain.2007.09.005 8 36 Ostelo RWJG, Deyo RA, Stratford P, et al. Interpreting change scores for pain and 9 functional status in low back pain: Towards international consensus regarding minimal 10 important change. Spine 2008;33:90–4. doi:10.1097/BRS.0b013e31815e3a10 11 37 Busse JW, Bartlett SJ, Dougados M, et al. Optimal strategies for reporting pain in 12 Confidential: For Review Only clinical trials and systematic reviews: Recommendations from an OMERACT 12 13 14 workshop. J Rheumatol 2015;42:1962–70. doi:10.3899/jrheum.141440 15 38 Ferreira ML, Herbert RD, Ferreira PH, et al. The smallest worthwhile effect of 16 nonsteroidal anti-inflammatory drugs and physiotherapy for chronic low back pain: A 17 benefit-harm trade-off study. J Clin Epidemiol 2013;66:1397–404. 18 doi:10.1016/j.jclinepi.2013.02.018 19 39 Chou R, Deyo R, Friedly J, et al. Systemic pharmacologic therapies for low back pain: 20 21 A systematic review for an American College of physicians clinical practice guideline. 22 Ann Intern Med 2017;166:480–92. doi:10.7326/M16-2458 23 40 Harrer M, Cuijpers P, Furukawa TA, et al. Doing meta-analysis in R: A hands-on 24 guide. Prot Lab Erlangen 2019. 25 41 Prescriber’s Digital Reference. PDR LLC. 2020.https://www.pdr.net 26 42 MIMS: Monthly Index of Medical Specialities. 2020.https://www.mims.co.uk/ 27 43 Australian Medicines Handbook 2020 (online). Adelaide: Australian Medicines 28 29 Handbook Pty Ltd. 2020.https://amhonline.amh.net.au/ 30 44 Akhter N, Siddiq MZ. Comparative efficacy of diclofenac sodium alone and in 31 combination with thiocolchicoside in patients with low back pain. Med Forum 32 2017;28:93–6. 33 45 Aksoy C, Karan A, Diraçoǧlu D. Low back pain: Results of an open clinical trial 34 comparing the standard treatment alone to the combination of standard treatment and 35 thiocolchicoside. J Orthop Traumatol 2002;3:103–8. doi:10.1007/s101950200036 36 37 46 Emrich OMD, Milachowski KA, Strohmeier M. Methocarbamol bei akuten 38 Rückenschmerzen: Eine randomisierte, doppelblinde, placebokontrollierte Studie. 39 MMW-Fortschritte der Medizin 2015;157:9–16. doi:10.1007/s15006-015-3307-x 40 47 Fathie K. A second look at skeletal muscle relaxant: a double-blind study with 41 metaxalone. Curr Ther Res 1964;6:677–83. 42 48 Foster L, Clapp L, Erickson M, et al. Botulinum toxin A and chronic low back pain a 43 44 randomized, double-blind study. Neurology 2001;56:1290–3. 45 doi:10.1212/WNL.56.10.1290 46 49 Friedman BW, Dym AA, Davitt M, et al. Naproxen with cyclobenzaprine, 47 oxycodone/acetaminophen, or placebo for treating acute low back pain: A randomized 48 clinical trial. JAMA 2015;314:1572–80. doi:10.1001/jama.2015.13043 49 50 Friedman BW, Irizarry E, Solorzano C, et al. Diazepam Is No Better Than Placebo 50 When Added to Naproxen for Acute Low Back Pain. Ann Emerg Med 2017;70:169- 51 52 176.e1. doi:10.1016/j.annemergmed.2016.10.002 53 51 Friedman BW, Cisewski D, Irizarry E, et al. A Randomized, Double-Blind, Placebo- 54 Controlled Trial of Naproxen With or Without Orphenadrine or Methocarbamol for 55 Acute Low Back Pain. Ann Emerg Med 2018;71:348-356.e5. 56 doi:10.1016/j.annemergmed.2017.09.031 57 52 Friedman BW, Irizarry E, Solorzano C, et al. A Randomized, Placebo-Controlled Trial 58 of Ibuprofen Plus Metaxalone, Tizanidine, or Baclofen for Acute Low Back Pain. Ann 59 60 Emerg Med 2019;74:512–20. doi:10.1016/j.annemergmed.2019.02.017

Cashin et al. 2021 Page 21 of 31 https://mc.manuscriptcentral.com/bmj Page 23 of 74 BMJ

1 2 3 53 Goforth HW, Preud’homme XA, Krystal AD. A Randomized, Double-Blind, Placebo- 4 Controlled Trial of Eszopiclone for the Treatment of Insomnia in Patients with 5 6 Chronic Low Back Pain. Sleep 2014;37:1053–60. doi:10.5665/sleep.3760 7 54 Gold RH. Orphenadrine citrate: sedative or muscle relaxant? Clin Ther 1978;1:451–3. 8 55 Herskowitz A. BOTOX (Botulinum Toxin Type A) treatment of patients with sub- 9 acute low back pain: A randomized, double blind, placebo-controlled study. J Pain 10 2004;5:S62. doi:10.1016/j.jpain.2004.02.214 11 56 Aparna P, Geetha P, Shanmugasundaram P. Comparison of aceclofenac and 12 Confidential: For Review Only combination (Aceclofenac + thiocolchicoside) therapy in acute low back pain patients. 13 14 Res J Pharm Technol 2016;9:1927–9. doi:10.5958/0974-360X.2016.00394.2 15 57 Hindle T, Palma L. Comparison of carisoprodol, butabarbital, and placebo in treatment 16 of low back syndrome. Calif Med 1972;117:7–11. 17 58 Hingorani K. Diazepam in backache: A double-blind controlled trial. Ann Phys Med 18 1966;8:303–6. 19 59 Jazayeri SM, Ashraf A, Fini HM, et al. Efficacy of Botulinum Toxin Type A for 20 21 Treating Chronic Low Back Pain. Anesthesiol Pain Med 2011;1:77–80. 22 doi:10.5812/kowsar.22287523.1845 23 60 Ketenci A, Ozcan E, Karamursel S. Assessment of efficacy and psychomotor 24 performances of thiocolchicoside and tizanidine in patients with acute low back pain. 25 Int J Clin Pract 2005;59:764–70. doi:10.1111/j.1742-1241.2004.00454.x 26 61 Klinger N., Wilson R., Kanniainen C., et al. Intravenous oprhenadrine for the 27 treatment of lumbar paravertebral muscle strain. Curr Ther Res 1988;43:247–54. 28 29 62 Lepisto P. A Comparative Trial of DS 103-282 and Placebo in the Treatment of Acute 30 Skeletal Muscle Spasms Due to Disorders of the Back. Curr Ther Res 1979;26:454–9. 31 63 Machado D, Kumar A, Jabbari B. Abobotulinum toxin A in the treatment of chronic 32 low back pain. Toxins 2016;8. doi:10.3390/toxins8120374 33 64 Pareek A, Chandurkar N, Chandanwale AS, et al. Aceclofenac-tizanidine in the 34 treatment of acute low back pain: A double-blind, double-dummy, randomized, 35 multicentric, comparative study against aceclofenac alone. Eur Spine J 2009;18:1836– 36 37 42. doi:10.1007/s00586-009-1019-4 38 65 Ralph L, Look M, Wheeler W, et al. Double-blind, placebo-controlled trial of 39 carisoprodol 250-mg tablets in the treatment of acute lower-back spasm. Curr Med Res 40 Opin 2008;24:551–8. doi:10.1185/030079908X261014 41 66 Salvini S, Antonelli S, De Micheli G, et al. Dantrolene sodium in low back pain and 42 cervico brachialgia treatment: a controlled study. Curr Ther Res 1986;39:172–7. 43 44 67 Baratta RR. A double-blind study of cyclobenzaprine and placebo in the treatment of 45 acute musculoskeletal conditions of the low back. Curr Ther Res 1982;32:646–52. 46 68 Schliessbach J, Vuilleumier PH, Siegenthaler A, et al. Analgesic effect of clobazam in 47 chronic low-back pain but not in experimentally induced pain. Eur J Pain 48 2017;21:1336–45. doi:10.1002/ejp.1032 49 69 Serfer GT, Wheeler WJ, Sacks HJ. Randomized, double-blind trial of carisoprodol 250 50 mg compared with placebo and carisoprodol 350 mg for the treatment of low back 51 52 spasm. Curr Med Res Opin 2010;26:91–9. doi:10.1185/03007990903382428 53 70 Tervo T, Petaja L, Lepisto P. A controlled clinical trial of a muscle relaxant analgesic 54 combination in the treatment of acute lumbago. Br J Clin Pract 1976;30:62–4. 55 71 Thompson M, Kennedy G. Treatment of acute low back pain: compartive trial of two 56 muscle relaxants, tizanidine and chlormezanone with placebo. Scandinavian Journal of 57 Rheumatology. 1983. 4–40. doi:10.3109/03009748309118006 58 72 Tüzün F, Ünalan H, Öner N, et al. Multicenter, randomized, double-blinded, placebo- 59 60 controlled trial of thiocolchicoside in acute low back pain. Jt Bone Spine 2003;70:356–

Cashin et al. 2021 Page 22 of 31 https://mc.manuscriptcentral.com/bmj BMJ Page 24 of 74

1 2 3 61. doi:10.1016/S1297-319X(03)00075-7 4 73 Zaringhalam J, Manaheji H, Rastqar A, et al. Reduction of chronic non-specific low 5 6 back pain: A randomised controlled clinical trial on acupuncture and baclofen. Chin 7 Med 2010;5:1–7. doi:10.1186/1749-8546-5-15 8 74 Berry H, Hutchinson DR. A Multicentre Placebo-Controlled Study in General Practice 9 to Evaluate the Efficacy and Safety of Tizanidine in Acute Low-Back Pain. J Int Med 10 Res 1988;16:75–82. doi:10.1177/030006058801600201 11 75 Berry H, Hutchinson DR. Tizanidine and Ibuprofen in Acute Low-Back Pain: Results 12 Confidential: For Review Only of a Double-Blind Multicentre Study in General Practice. J Int Med Res 1988;16:83– 13 14 91. doi:10.1177/030006058801600202 15 76 Borenstein DG, Lacks S, Wiesel SW. Cyclobenzaprine and naproxen versus naproxen 16 alone in the treatment of acute low back pain and muscle spasm. Clin Ther 17 1990;12:125–31. 18 77 Casale R. Acute low back pain: Symptomatic treatment with a muscle relaxant drug. 19 Clin J Pain 1988;4:81–8. 20 21 78 Cogné M, Petit H, Creuzé A, et al. Are paraspinous intramuscular injections of 22 botulinum toxin a (BoNT-A) efficient in the treatment of chronic low-back pain? A 23 randomised, double-blinded crossover trial. BMC Musculoskelet Disord 2017;18:454. 24 doi:10.1186/s12891-017-1816-6 25 79 Dapas F, Hartman SF, Martinez L, et al. Baclofen for the treatment of acute low-back 26 syndrome: A double-blind comparison with placebo. Spine 1985;10:345–9. 27 doi:10.1097/00007632-198505000-00010 28 29 80 ACTRN12616000017426. A randomised controlled feasibility study of managing 30 sleep with Zopiclone in participants with acute low back pain and sleep disturbances. 31 Aust. New Zeal. Clin. Trials Regist. 2016. 32 81 EUCTR2017-004530-29. No Title. Clinicaltrialsregister.eu. 33 2017.https://www.clinicaltrialsregister.eu/ctr-search/trial/2017-004530-29/HU 34 82 NCT02887534. Evaluation of Efficacy and Safety of SPARC1401 in Acute Low Back 35 Pain. ClinicalTrials.gov. 2016. 36 37 83 NCT01587508. Study Comparing A New Drug Containing The Combination 38 Meloxicam And Cyclobenzaprine In The Treatment Of Acute Lumbago. 39 ClinicalTrials.gov. 2012. 40 84 EUCTR2019-001885-14. No Title. Clinicaltrialsregister.eu. 41 2019.https://www.clinicaltrialsregister.eu/ctr-search/trial/2019-001885-14/HU 42 85 IRCT20111109008035N4. The study of efficacy of melatonin, and zolpidem on the 43 44 sleep quality, and severity of pain in the patients with chronic non-specific low back 45 pain. ISRCTN. 2017. 46 86 NCT00671879. Study to Evaluate Two Formulations of Carisoprodol in Subjects With 47 Musculoskeletal Spasm of the Lower Back. ClinicalTrials.gov. 48 2008.https://clinicaltrials.gov/ct2/show/NCT00671879 49 87 NCT00671502. A Study to Evaluate Two Formulations of Carisoprodol in Subjects 50 With Musculoskeletal Spasm of the Lower Back. ClinicalTrials.gov. 2008. 51 52 88 NCT00817986. A Study to Evaluate the Safety and Tolerability of Arbaclofen 53 Placarbil (XP19986) in Subjects With Acute Back Spasms. ClinicalTrials.gov. 2009. 54 89 NCT00404417. Botulinum Toxin A for the Treatment of Chronic Lumbar Back Pain. 55 ClinicalTrials.gov. 2006. 56 90 NCT00384579. Pilot Study to Assess the Efficacy of Botulinum Toxin B on Pain and 57 Disability in Subjects With Acute Low Back Pain. ClinicalTrials.gov. 2006. 58 91 NCT00384371. Pilot Study to Assess the Efficacy of Botulinum Toxin A Treatments 59 60 on Pain and Disability in Sub-Acute Low Back Pain. ClinicalTrials.gov. 2006.

Cashin et al. 2021 Page 23 of 31 https://mc.manuscriptcentral.com/bmj Page 25 of 74 BMJ

1 2 3 92 van Tulder MW, Touray T, Furlan AD, et al. Muscle relaxants for non-specific low- 4 back pain. Cochrane Database Syst Rev 2003;2017. doi:10.1002/14651858.CD004252 5 6 93 Michaleff ZA, Harrison C, Britt H, et al. Ten-year survey reveals differences in GP 7 management of neck and back pain. Eur Spine J 2012;21:1283–9. doi:10.1007/s00586- 8 011-2135-5 9 94 Bramness JG, Buajordet I, Skurtveit S. The role of pharmacoepidemiological studies 10 in the market withdrawal of carisoprodol (Somadril®) in Europe. Nor Epidemiol 11 2008;18:167–72. doi:10.5324/nje.v18i2.29 12 Confidential: For Review Only 95 Carisoprodol and meprobamate: risks outweigh benefits - GOV.UK. 13 14 https://www.gov.uk/drug-safety-update/carisoprodol-and-meprobamate-risks- 15 outweigh-benefits (accessed 10 Mar 2021). 16 96 Underwood M, Tysall C. Antidepressants for musculoskeletal pain. BMJ 2021;372. 17 doi:10.1136/bmj.n80 18 97 Henschke N, Maher CG, Refshauge KM, et al. Prognosis in patients with recent onset 19 low back pain in Australian primary care: inception cohort study. BMJ 20 21 2008;337:a171–a171. doi:10.1136/bmj.a171 22 98 Witenko C, Moorman-Li R, Motycka C, et al. Considerations for the appropriate use 23 of skeletal muscle relaxants for the management of acute low back pain. P T 24 2014;39:427–35. 25 99 Chiarotto A, Deyo RA, Terwee CB, et al. Core outcome domains for clinical trials in 26 non-specific low back pain. Eur Spine J 2015;24:1127–42. doi:10.1007/s00586-015- 27 3892-3 28 29 100 Dworkin RH, Turk DC, Farrar JT, et al. Core outcome measures for chronic pain 30 clinical trials: IMMPACT recommendations. Pain 2005;113:9–19. 31 doi:10.1016/j.pain.2004.09.012 32 101 Schulz KF, Altman DG, Moher D. CONSORT 2010 Statement: updated guidelines for 33 reporting parallel group randomised trials. BMJ 2010;340:c332–c332. 34 doi:10.1136/bmj.c332 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 24 of 31 https://mc.manuscriptcentral.com/bmj BMJ Page 26 of 74

1 2 3 Table 1. Muscle relaxant medicine primer. 4 5 6 Group Clinical utility Medicine examples 7 included in this review 8 Antispastic To reduce heightened muscle tone Baclofen, dantrolene 9 (spasticity) commonly associated 10 with cerebral palsy, multiple sclerosis 11 and spinal cord injuries. 12 Confidential: For Review Only 13 Non- To reduce acute muscle spasm Carisoprodol, 14 benzodiazepine commonly associated with muscle cyclobenzaprine, 15 antispasmodic injury. These medicines also have a metaxalone, methocarbamol, 16 strong sedative action. thiocolchicoside, tizanidine, 17 tolperisone, orphenadrine 18 Benzodiazepine To reduce acute muscle spasm Diazepam 19 20 commonly associated with muscle 21 injury. These medicines also have a 22 strong sedative action as well as 23 anxiolytic, hypnotic, and 24 anticonvulsant actions. 25 Miscellaneous Although less commonly classified as Botulinum toxin, 26 27 muscle relaxants, several other eszopiclone 28 medicines are prescribed for their 29 ability to reduce muscle spasm and/or 30 muscle tone (spasticity). These 31 include botulinum toxins and non- 32 benzodiazepine hypnotics. 33 34 35 Muscle relaxant medicines are generally prescribed to reduce muscle spasm and/or muscle 36 tone (spasticity). The term “muscle relaxant” is very broad and includes many different 37 chemically unrelated medicines with different clinical utility and mechanisms of action.[21] 38 The choice of muscle relaxant medicine and frequency of prescription by a physician varies 39 between countries,[9,93] with considerable clinical uncertainty in preferencing one muscle 40 relaxant medicine over another. The muscle relaxant medicines have been grouped according 41 to their clinical utility. 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 25 of 31 https://mc.manuscriptcentral.com/bmj Page 27 of 74 BMJ

1 2 3 Table 2. Summary of findings and certainty of evidence for pain 4 5 Summary of findings Certainty of evidence 6 No of participants Mean difference Risk of bias Inconsistency Imprecision Publication bias Certainty 7 (No of trials) (95% CI), 0-100 of 8 Confidential: For Review Only evidence 9 Acute LBP 10 Non-benzodiazepine antispasmodics: 11 ≤ 2 weeks 4546 (16) -7.7 (-12.1 to -3.3) Downgradeda Downgradedc Downgradedd Not downgraded Very low 12 3-13 weeks 612 (3) 0.6 (-4.5 to 5.7) Not Not downgraded Downgradedd Not downgraded Moderate 13 downgraded 14 Antispastics: 15 ≤ 2 weeks 103 (1) -1.6 (-15.3 to 12.1) Not Not downgradedf Downgradede Not downgradedf Low 16 downgraded 17 3-13 weeks 99 (1) 4 (-7.7 to 15.7) Not Not downgradedf Downgradedd Not downgradedf Moderate 18 downgraded 19 Benzodiazepines: 20 f d f 21 ≤ 2 weeks 112 (1) 2 (-9.8 to 13.8) Not Not downgraded Downgraded Not downgraded Moderate 22 downgraded f e f 23 3-13 weeks 103 (1) -1 (-10.4 to 8.4) Not Not downgraded Downgraded Not downgraded Low 24 downgraded 25 Subacute LBP 26 Miscellaneous: 27 3-13 weeks 28 (1) -19 (-41.9 to 3.9) Downgradeda Not downgradedf Downgradede Not downgradedf Very low 28 Chronic LBP 29 Antispastics: 30 3-13 weeks 80 (1) -5.4 (-13.7 to 2.9) Downgradeda Not downgraded Downgradede Not downgraded Very low 31 Miscellaneous: 32 3-13 weeks 52 (1) -19.9 (-31.5 to -8.3) Not Not downgradedf Downgradedd Not downgradedf Moderate 33 downgraded 34 Mixed LBP 35 Non-benzodiazepine antispasmodics: 36 ≤ 2 weeks 617 (2) -4.4 (-6.9 to -1.9) Downgradeda Not downgraded Not downgraded Not downgraded Low 37 3-13 weeks 329 (1) -5.8 (-13.8 to 2.2) Downgradeda Not downgradedf Downgradede Not downgradedf Very low 38 39 LBP, Low Back Pain; CI, Confidence Interval a 40 Downgraded by two levels because >50% of participants in this comparison were from studies at high risk of bias 41 42 Cashin et al. 2021 Page 26 of 31 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 28 of 74

1 2 3 bDowngraded by one level because >25% but <50% of participants in this comparison were from studies at high risk of bias 4 cDowngraded by one level because heterogeneity (I2) > 50% dDowngraded by one level because the limits of the 95% CI crosses the minimally clinically important difference or the null 5 eDowngraded by two levels because the limits of the 95% CI crosses the minimally clinically important difference and the null 6 fNot downgraded because can’t be determined with one study 7 8 Confidential: For Review Only 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 Cashin et al. 2021 Page 27 of 31 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 29 of 74 BMJ

1 2 3 Table 3. Summary of findings and certainty of evidence for disability 4 5 Summary of findings Certainty of evidence 6 No of Mean difference (95% CI), Risk of bias Inconsistency Imprecision Publication bias Certainty of 7 participants 0-100 evidence 8 (No ofConfidential: trials) For Review Only 9 Acute LBP 10 Non-benzodiazepine antispasmodics: 11 ≤ 2 weeks 2438 (7) -3.3 (-7.3 to 0.7) Downgradeda Downgradedc Downgradedd Not downgraded Very low 12 3-13 weeks 422 (2) 4.3 (-1.4 to 10.1) Not downgraded Not downgraded Downgradedd Not downgraded Moderate 13 Antispastics: 14 ≤ 2 weeks 103 (1) 2 (-15.6 to 19.6) Not downgraded Not downgradedf Downgradede Not downgradedf Low 15 Benzodiazepines: 16 ≤ 2 weeks 112 (1) 0 (-13.2 to 13.2) Not downgraded Not downgradedf Downgradede Not downgradedf Low 17 f d f 18 3-13 weeks 103 (1) -6.9 (-12.1 to -1.7) Not downgraded Not downgraded Downgraded Not downgraded Moderate 19 Chronic 20 Antispastics: a d 21 3-13 weeks 80 (1) -3.2 (-8.3 to 1.8) Downgraded Not downgraded Downgraded Not downgraded Very low 22 Miscellaneous: 23 3-13 weeks 52 (1) -5.6 (-20.6 to 9.4) Not downgraded Not downgradedf Downgradede Not downgradedf Low 24 Mixed LBP 25 Non-benzodiazepine antispasmodics: 26 ≤ 2 weeks 329 (1) -19.2 (-27.7 to -10.7) Downgradeda Not downgradedf Not downgraded Not downgradedf Low 27 LBP, Low Back Pain; CI, Confidence Interval 28 aDowngraded by two levels because >50% of participants in this comparison were from studies at high risk of bias 29 bDowngraded by one level because >25% but <50% of participants in this comparison were from studies at high risk of bias 30 cDowngraded by one level because heterogeneity (I2) > 50% d 31 Downgraded by one level because the limits of the 95% CI crosses the minimally clinically important difference or the null eDowngraded by two levels because the limits of the 95% CI crosses the minimally clinically important difference and the null 32 fNot downgraded because can’t be determined with one study 33 34 35 36 37 38 39 40 41 42 Cashin et al. 2021 Page 28 of 31 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 30 of 74

1 2 3 Table 4. Summary of findings and certainty of evidence for acceptability 4 5 Summary of findings Certainty of evidence 6 No of Relative risk (95% CI) Risk of bias Inconsistency Imprecision Publication bias Certainty of 7 participants evidence 8 (NoConfidential: of trials) For Review Only 9 Acute LBP 10 Non- 2834 (13) 0.8 (0.6 to 1.1) Downgradeda Not downgraded Downgradedb Not downgraded Very low 11 benzodiazepine 12 antispasmodics 13 Chronic LBP 14 Antispastics 84 (1) 1.6 (0.2 to 12.9) Downgradeda Not downgraded Downgradedb Not downgradedc Very low 15 Miscellaneous 101 (2) 0.6 (0.2 to 1.7) Not downgraded Not downgraded Downgradedb Not downgraded Moderate 16 17 LBP, Low Back Pain; CI, Confidence Interval 18 aDowngraded by two levels because >50% of participants in this comparison were from studies at high risk of bias bDowngraded by one level because the limits of the 95% CI crosses the null 19 cNot downgraded because can’t be determined with one study 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 Cashin et al. 2021 Page 29 of 31 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 31 of 74 BMJ

1 2 3 Table 5. Summary of findings and certainty of evidence for safety 4 5 Summary of findings Certainty of evidence 6 No of Risk difference (95% CI) Risk of bias Inconsistency Imprecision Publication bias Certainty of 7 participants evidence 8 Confidential:(No of trials) For Review Only 9 Adverse events: 10 Acute LBP 11 Non-benzodiazepine 3404 (16) 1.6 (1.2 to 2) Downgradeda Not downgraded Not downgraded Not downgraded Low 12 antispasmodics 13 Antispastic 290 (2) 2 (1.1 to 3.8) Downgradedb Not downgraded Not downgraded Not downgraded Moderate 14 Benzodiazepines 159 (2) 1.8 (0.9 to 3.6) Downgradedb Not downgraded Downgradedc Not downgraded Low 15 Chronic LBP 16 Miscellaneous 95 (2) 1.5 (0.4 to 5.7) Not Downgraded Not downgraded Downgradedc Not downgraded Moderate 17 Mixed LBP 18 a d c d 19 Non-benzodiazepine 329 (1) 1.6 (0.6 to 4.3) Downgraded Not downgraded Downgraded Not downgraded Very low 20 antispasmodics 21 Serious adverse events: 22 Acute LBP a c 23 Non-benzodiazepine 830 (2) 2.3 (0.3 to 20.8) Downgraded Not downgraded Downgraded Not downgraded Very low 24 antispasmodics 25 Tolerability 26 Acute LBP 27 Non-benzodiazepine 1641 (5) 1.5 (0.6 to 3.5) Downgradeda Not downgraded Downgradedc Not downgraded Very low 28 antispasmodics 29 Antispastic 195 (1) 34.6 (2.1 to 568) Downgradeda Not downgradedd Downgradedc Not downgraded Very low 30 LBP, Low Back Pain; CI, Confidence Interval 31 aDowngraded by two levels because >50% of participants in this comparison were from studies at high risk of bias 32 bDowngraded by one level because >25% but <50% of participants in this comparison were from studies at high risk of bias 33 cDowngraded by one level because the limits of the 95% CI crosses the null d 34 Not downgraded because can’t be determined with one study 35 36 37 38 39 40 41 42 Cashin et al. 2021 Page 30 of 31 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 32 of 74

1 2 3 Figure captions 4 5 Figure 1. PRISMA flow diagram of the record selection process. 6 7 8 Figure 2. Effect of non-benzodiazepine antispasmodic medicines compared to control on 9 10 pain intensity (0-100 scale) at immediate term (≤ 2 weeks) post randomisation for adults with 11 12 LBP. NegativeConfidential: values for the mean difference For indicate Review that the effect Onlyfavours muscle relaxant 13 14 medicines compared to control. Whereas, negative values for the trial observations indicate 15 change from baseline. 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 31 of 31 https://mc.manuscriptcentral.com/bmj Page 33 of 74 BMJ

1 2 3 Supplementary Online Content 4 5 6 Supplemental file 1. Deviations from protocol 7 8 Supplemental file 2. Search strategy Ovid MEDLINE 9 10 Supplemental file 3. Search strategies for trial registries 11 12 SupplementalConfidential: file 4. Interventions of interest For Review Only 13 14 15 Supplemental file 5. GRADE framework 16 17 Supplemental file 6. Calculation of effect sizes for pain intensity 18 19 Supplemental file 7. Calculation of effect sizes for disability 20 21 Supplemental file 8. Characteristics of included studies 22 23 Supplemental file 9. Narrative description of trials not included in meta-analysis for pain intensity 24 25 Supplemental file 10. Risk of bias assessments 26 27 Supplemental file 11. Forest plot pain intensity 3-13 weeks 28 29 Supplemental file 12. Forest plot disability ≤ 2 weeks 30 31 Supplemental file 13. Forest plot disability 3-13 weeks 32 33 Supplemental file 14. Forest plot acceptability 34 35 36 Supplemental file 15. Forest plot adverse events 37 38 Supplemental file 16. Forest plot serious adverse events 39 40 Supplemental file 17. Forest plot tolerability 41 42 Supplemental file 18. Forest plot dose subgroup analysis 43 44 Supplemental file 19. Sensitivity analyses for non-benzodiazepine antispasmodic medicines in acute 45 46 LBP 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 1 of 41 https://mc.manuscriptcentral.com/bmj BMJ Page 34 of 74

1 2 3 Supplemental file 1. Deviations from protocol 4 5 6 We deviated from our pre-registered protocol (accessed from https://osf.io/mu2f5/) to 7 improve both the clinical interpretability and comparability of the review findings. 8 9 The deviations are as follows: 10 11 • We redefined the follow-up timepoints in relation to ‘post-randomisation’ as opposed 12 toConfidential: ‘post-treatment’ to ensure comparable For followReview-up between trials.Only The follow-up 13 timepoints are now immediate (≤ 2 weeks) and short-term (3-13 weeks). 14 15 • We redefined how the muscle relaxant medicines were grouped to better reflect 16 clinical utility from (antispasmodic or antispastic) to (non-benzodiazepine 17 antispasmodic, antispastic, benzodiazepine and miscellaneous). 18 19 • We conducted additional ad hoc sensitivity analyses investigating the effect of 20 removing trials at high risk of bias, trials primarily reported as trial registry records, 21 trials without a placebo comparison, and trials investigating the muscle relaxant 22 medicine carisoprodol. 23 24 • We did not report the extended funnel plot following reviewer recommendations. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 2 of 41 https://mc.manuscriptcentral.com/bmj Page 35 of 74 BMJ

1 2 3 Supplemental file 2. Search strategy Ovid MEDLINE 4 5 Search Strategy for Ovid MEDLINE: 6 7 Part A: Generic search for randomized controlled trials 8 1. randomized controlled trial.pt. 9 2. controlled clinical trial.pt. 10 3. comparative study.pt. 11 4. clinical trial.pt. 12 Confidential: For Review Only 5. random*.ab,ti. 13 14 6. placebo.ab,ti. 15 7. drug therapy.fs. 16 8. trial.ab,ti. 17 9. groups.ab,ti. 18 10. or/1-9 19 11. (animals not (humans and animals)).sh. 20 12. (adolescent* or teen* or youth? or puberty or childhood or children* or p?ediatri* or preschool or 21 pre-school or nursery or kindergarten or infant? or newborn? or neonat* or prematurity or fetal or 22 foetal).mp. 23 13. 11 or 12 24 25 14. 10 not 13 26 27 Part B: Specific search for low back, sacrum and coccyx problems 28 15. dorsalgia.ti,ab. 29 16. exp Back Pain/ 30 17. backache.ti,ab. 31 18. (lumbar adj pain).ti,ab. 32 19. coccydynia.ti,ab. 33 20. sciatica.ti,ab. 34 21. spondylosis.ti,ab. 35 36 22. lumbago.ti,ab. 37 23. back disorder$.ti,ab 38 24. or/15-23 39 40 Part C: Specific search for other spinal disorders 41 25. Coccyx.sh 42 26. Lumbar Vertebrae.sh 43 27. Intervertebral disc.sh 44 28. Sacrum.sh 45 29. Intervertebral disc degeneration.sh 46 47 30. (disc adj degeneration).ti,ab. 48 31. (disc adj prolapse).ti,ab. 49 32. (disc adj herniation).ti,ab. 50 33. spinal fusion.sh. 51 34. (facet adj joints).ti,ab. 52 35. Intervertebral Disc Displacement.sh. 53 36. or/25-35 54 55 Part D: Specific search for interventions of interest 56 37. suxamethonium.mp. or Succinylcholine/ 57 58 38. exp Botulinum Toxins/ 59 39. pancuronium/ 60 40. Vecuronium Bromide/

Cashin et al. 2021 Page 3 of 41 https://mc.manuscriptcentral.com/bmj BMJ Page 36 of 74

1 2 3 41. Atracurium/ 4 42. Rocuronium/ 5 43. mivacurium bromide.mp. 6 44. cisatracurium.mp. 7 8 45. Carisoprodol/ 9 46. Methocarbamol/ 10 47. Chlorzoxazone/ 11 48. Orphenadrine/ 12 49. Baclofen/Confidential: For Review Only 13 50. tizanidine.mp. 14 51. Tolperisone/ 15 52. thiocolchicoside.mp. 16 53. cyclobenzaprine.mp. 17 54. Dantrolene/ 18 19 55. Clonazepam/ 20 56. exp Diazepam/ 21 57. Chlordiazepoxide/ 22 58. Oxazepam/ 23 59. Lorazepam/ 24 60. Bromazepam/ 25 61. Clobazam/ 26 62. Alprazolam/ 27 63. clotiazepam.mp. 28 64. Flurazepam/ 29 30 65. Nitrazepam/ 31 66. Flunitrazepam/ 32 67. Estazolam/ 33 68. Triazolam/ 34 69. lormetazepam.mp. 35 70. Temazepam/ 36 71. Midazolam/ 37 72. quazepam.mp. 38 73. Zolpidem/ 39 74. zaleplon.mp. 40 41 75. Eszopiclone/ 42 76. metaxalone.mp. 43 77. or/37-76 (all interventions of interest) 44 45 Results 46 78. 24 or 36 (all back pain) 47 79. 77 and 78 (all back pain and all interventions of interest) 48 80. 14 and 79 (all RCTs of interventions of interest in back pain) 49 50 51 52 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 4 of 41 https://mc.manuscriptcentral.com/bmj Page 37 of 74 BMJ

1 2 3 Supplemental file 3. Search strategies for trial registries 4

5 6 Muscle Relaxant Medicines 7 8 WHO ICTRP: Advanced 9 search 10 11 Title: – 12 Confidential: For Review Only 13 Condition: ‘back pain’ 14 15 Intervention: 1-40 16 17 Recruitment status: ALL 18 19 Phases are: ALL 20 21 ClinicalTrials.gov: 22 Advanced search 23 24 Study Type: Interventional Studies 25 26 Study Results: All studies 27 28 Recruitment: All studies 29 30 Age: Adult and Senior 31 32 Gender: All studies 33 Conditions: ‘back pain’ 34 35 Interventions: 1-40 36 37 Titles: – 38 39 Outcome Measures: – 40 41 Sponsor/Collaborators: – 42 43 Sponsor (Lead): – 44 45 Study IDs: – 46 47 Locations: – 48 49 Phase: – 50 51 Funder Type: – 52 53 First Received: – 54 55 Last Updated: – 56 57 58 59 60

Cashin et al. 2021 Page 5 of 41 https://mc.manuscriptcentral.com/bmj BMJ Page 38 of 74

1 2 3 EU ClinicalTrials Register: Muscle Relaxant Medicines 4 Advanced search 5 6 Search Term: back pain AND ‘intervention’ 7 (1-40) 8 9 Country: – 10 11 Age Range: Adult and Elderly 12 Confidential: For Review Only 13 Trial Status: – 14 15 Trial Phase: – 16 17 Gender: Both 18 19 Date Range: – 20 21 Results Status: – 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 6 of 41 https://mc.manuscriptcentral.com/bmj Page 39 of 74 BMJ

1 2 3 Supplemental file 4. Interventions of interest 4

5

6 Drug name ATC code Licenses 7 8 Number ARTG FDA EMA 9 1 10 suxamethonium M03AB01 yes - yes 11 2 botulinum toxin M03AX01 yes yes yes 12 Confidential: For Review Only 3 pancuronium M03AC01 yes yes - 13 4 14 vecuronium M03AC03 yes yes yes 15 5 atracurium M03AC04 - yes - 16 6 rocuronium bromide M03AC09 - - yes 17 7 18 mivacurium bromide M03AC10 yes - yes 19 8 cisatracurium M03AC11 yes yes yes 20 9 M03BA02 21 carisoprodol - yes yes 22 10 methocarbamol M03BA03 - yes - 23 11 chlorzoxazone M03BB03 - yes - 24 12 M03BC01 25 orphenadrine citrate yes yes - 26 13 baclofen M03BX01 yes yes yes 27 14 tizanidine M03BX02 - yes yes 28 15 M03BX04 29 tolperisone - - yes 30 16 thiocolchicoside M03BX05 - - yes 31 17 cyclobenzaprine M03BX08 - yes - 32 18 33 dantrolene M03CA01 yes yes yes 34 19 clonazepam N03AE01 yes yes yes 35 20 diazepam N05BA01 yes yes - 36 21 37 chlordiazepoxide N05BA02 - yes - 38 22 oxazepam N05BA04 yes yes - 39 23 lorazepam N05BA06 yes yes yes 40 24 41 bromazepam N05BA08 yes - yes 42 25 clobazam N05BA09 yes yes - 43 26 N05BA12 44 alprazolam yes yes yes 45 27 clotiazepam N05BA21 - - yes 46 28 flurazepam N05CD01 - yes - 47 29 N05CD02 48 nitrazepam yes - yes 49 30 flunitrazepam N05CD03 yes - yes 50 31 estazolam N05CD04 - yes - 51 32 N05CD05 52 triazolam yes yes yes 53 33 lormetazepam N05CD06 - - yes 54 34 temazepam N05CD07 yes yes - 55 35 56 midazolam N05CD08 yes yes yes 57 36 quazepam N05CD10 - yes - 58 37 zolpidem N05CF02 yes yes - 59 38 60 zaleplon N05CF03 yes yes -

Cashin et al. 2021 Page 7 of 41 https://mc.manuscriptcentral.com/bmj BMJ Page 40 of 74

1 2 3 Licenses Drug name ATC code 4 5 Number ARTG FDA EMA 6 39 7 eszopiclone N05CF04 yes yes - 8 40 metaxalone - - yes - 9 10 11 12 Confidential: For Review Only 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 8 of 41 https://mc.manuscriptcentral.com/bmj Page 41 of 74 BMJ

1 2 3 Supplemental file 5. GRADE framework 4 5 6 Certainty in the evidence was assessed using the Grading of Recommendations Assessment, 1 7 Development and Evaluation (GRADE) working group methodology. The certainty of 8 evidence was initially classified as ‘high’ (very certain that the true effect lies close to that of 9 the estimate of the effect) and possibly downgraded to ‘moderate’ (moderately certain in the 10 effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a 11 12 possibilityConfidential: that it is substantially different ),For ‘low’ (certainty Review in the effect Only estimate is limited: The 13 true effect may be substantially different from the estimate of the effect), or ‘very low’ (very 14 little certainty in the effect estimate: The true effect is likely to be substantially different from 15 the estimate of effect). 16

17 18 We graded the evidence in the following recommended domains in the following manner: 19 • Risk of bias: we downgraded by one level if > 25% but < 50% of the participants in our 20 analysis came from trials assessed as ‘high’ risk of bias, and we downgraded by two 21 levels if > 50% of the patients came from trials assessed as ‘high’ risk of bias.2 22 23 • Inconsistency: we downgraded by one level if we identified important heterogeneity. 2 24 We assessed heterogeneity using the between-study variance parameter (τ ) and the 25 proportion of study variance not due to sampling error (I2).3 26 • Indirectness: we did not consider this domain because the eligibility criteria ensures 27 4 28 patients, interventions, and comparators were similar across studies. 29 • Imprecision: we downgraded by one level if the width of the confidence intervals (for 30 continuous variables as pain intensity and disability) by crossing either the null or the 31 threshold for a clinically meaningful effect (10 points on a 0 to 100 scale) and two levels 32 if the interval spanned both. For dichotomous variables (like harms) we downgraded 33 5 34 by one level if the interval spanned the null. 35 • Publication bias: we downgraded by only one level if we strongly detected publication 36 bias. We assessed publication bias by visually assessing funnel plot and sensitivity 37 analysis.6 38 39 40 References 41 42 1. Balshem H, Helfand M, Schunemann HJ, et al. GRADE guidelines: 3. Rating the quality of 43 evidence. J Clin Epidemiol. 2011;64(4):401-406 44 2. Guyatt GH, Oxman AD, Vist G, et al. GRADE guidelines: 4. Rating the quality of evidence 45 - study 3limitations (risk of bias). J Clin Epidemiol. 2011;64(4):407-415. 46 3. Guyatt GH, Oxman AD, Kunz R, et al. GRADE guidelines: 7. Rating the quality of evidence 47 - inconsistency. J Clin Epidemiol. 2011;64(12):1294-1302. 48 4. Guyatt GH, Oxman AD, Kunz R, et al. GRADE guidelines: 8. Rating the quality of evidence 49 - indirectness. J Clin Epidemiol. 2011;64(12):1303-1310. 50 5. Guyatt GH, Oxman AD, Kunz R, et al. GRADE guidelines 6. Rating the quality of evidence 51 - imprecision. J Clin Epidemiol. 2011;64(12):1283-1293. 52 6. Guyatt GH, Oxman AD, Montori V, et al. GRADE guidelines: 5. Rating the quality of 53 evidence - publication bias. J Clin Epidemiol. 2011;64(12):1277-1282 54 55 56 57 58 59 60

Cashin et al. 2021 Page 9 of 41 https://mc.manuscriptcentral.com/bmj BMJ Page 42 of 74

1 2 3 Supplementary file 6. Calculation of effect sizes for pain intensity 4 5 Author, year Muscle Outcome Type of Type of Point estimate Mean (SD), converteda Number of participants 6 relaxant scale data measure (variability) extracted 7 medicine extracted Muscle Comparator Muscle Comparator Muscle Comparato 8 Confidential: ForRelaxant ReviewRelaxant Only Relaxant r 9 Immediate term (≤ 2 weeks) 10 Acute LBP 11 Aparna 2016 Thiocolchicoside 0-10 VAS Mean FV 0.7 1.15 6.7 (30)b 11.5 (30)b 79 74 c 12 Baratta 1982 Cyclobenzaprine 0-10 VAS Mean (p- CS -5.5 -5 -55 -40 (48.9) 58 59 value) (48.9)c 13 Friedman 2015 Cyclobenzaprine 0-10 VAS Mean FV 3.6 3.9 36 (35.8)e 39 (30.9)e 103 104 14 (95% CI) 15 Friedman 2017 Diazepam VRS-4 Mean (SD) FV 1 (1) 0.9 (1) 31.7 29.7 (32) 57 55 16 (31.7) 17 Friedman 2018 Orphenadrine VRS-4 Mean (SD) FV 1.1 (1) 1.2(1) 38 (33) 39 (32) 78 38f 18 Friedman 2018 Methocarbamol VRS-4 Mean (SD) FV 1.3 (1) 1.2(1) 43 (32.7) 39 (32) 80 38f 19 Friedman 2019 Baclofen VRS-4 Mean (SD) FV 1.1 (1) 1.2 (0.9) 37.7 (32) 38.3 (29.3) 79 24f f 20 Friedman 2019 Metaxalone VRS-4 Mean (SD) FV 1.3 (1) 1.2 (0.9) 42 (33) 38.3 (29.3) 76 24 f 21 Friedman 2019 Tizanidine VRS-4 Mean (SD) FV 1.2 (1) 1.2 (0.9) 38.7 38.3 (29.3) 76 25 (31.7) 22 Hindle 1972 Carisoprodol 0-100 VAS Mean FV 15.5 64 15.5 (30)b 64 (30)b 14 14 23 Lepisto 1979 Tizanidine VRS-4 Mean CS -1.5 -1.6 -51 (30)b -52.7 (30)b 15 15 24 Pareek 2009 Tizanidine 0-10 VAS Mean (SD) CS -5.9 (2.1) -4.4 (2.1) -58.8 -43.5 (20.6) 94 91 25 (21.4) 26 Ralplh 2008 Carisoprodol VRS-4 Mean (SE) CS -1.9 (0.2) -1.2 (0.2) -47 -30 (66.7)d 269 278 27 (19.5)d d f 28 Serfer 2010 Carisoprodol A VRS-5 Mean (SE) CS -1.8 (0.1) -1.4 (0.1) -44.5 -34.3 (44) 260 128 d 29 (48.4) Serfer 2010 Carisoprodol B VRS-5 Mean (SE) CS -1.8 (0.1) -1.4 (0.1) -44.5 -34.3 (44)d 251 128f 30 (47.5)d 31 NCT00671879 Carisoprodol A 0-100 VAS Mean (SE) CS -15.5 -15.2 (1.3) -15.5 -15.2 (21.4)d 271 132f 32 (1.3) (22.1)d 33 NCT00671879 Carisoprodol B 0-100 VAS Mean (SE) CS -16.4 -15.2 (1.3) -16.4 -15.2 (21.4)d 270 132f 34 (1.3) (21.4)d 35 NCT00671502 Carisoprodol A 0-100 VAS Mean CS -27.5 -28.6 -27.5 -28.6 (30)b 280 140f b 36 (30) NCT00671502 Carisoprodol B 0-100 VAS Mean CS -28 -28.6 -28 (30)b -28.6 (30)b 281 139f 37 Mixed LBP 38 Akhter 2017 Thiocolchicoside 0-10 VAS Mean (SE) FV 0.94 (0.1) 1.35 (0.1) 9.4 (11.5)d 13.5 (11.5)d 144 144 39 40 41 42 Cashin et al. 2021 Page 10 of 41 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 43 of 74 BMJ

1 2 3 Short term (3-13 weeks) 4 Acute LBP 5 Friedman 2015 Cyclobenzaprine 0-10 VAS Mean FV 0.6 (1) 0.7 (1.1) 19.3 24.3 (35.3) 108 107 6 (95% CI) (31.7) 7 Friedman 2017 Diazepam VRS-4 Mean (SD) FV 0.3 (0.7) 0.4 (0.8) 11.3 (23) 12.3 (25.7) 50 53 f 8 Friedman 2018 Orphenadrine VRS-4 Mean (SD) FV 0.6 (0.9) 0.7 (1) 21.3 (29) 22.7 (34.7) 70 34 Friedman 2018 MethocarbamolConfidential: VRS-4 Mean (SD) FV For0.7 (1) Review0.7 (1) 24.7 (32) Only22.7 (34.7) 70 34f 9 Friedman 2019 Baclofen VRS-4 Mean (SD) FV 0.6 (0.9) 0.4 (0.7) 18.3 (31) 14.3 (23) 76 23f 10 Friedman 2019 Metaxalone VRS-4 Mean (SD) FV 0.6 (0.9) 0.4 (0.7) 20 (31) 14.3 (23) 72 23f 11 Friedman 2019 Tizanidine VRS-4 Mean (SD) FV 0.6 (0.9) 0.4 (0.7) 19.7 14.3 (23) 70 24f 12 (29.3) 13 Sub-acute LBP 14 Herskowitz 2004 Botulinum toxin 0-10 VAS Mean (p- CS -2.2 -0.3 -22 -3 (32.1)c 13 15 15 A value) (29.8)c 16 17 SD, standard deviation; MD, mean difference; 95% CI, 95% confidence interval; FV, Final Value; CS, Change Score; VAS, Visual Analogue Scale; VRS-4, Verbal Rating Scale 4 levels; VRS-5, Verbal Rating Scale 5 levels 18 a Mean and variability measures divided by the top number of scale and multiplied by 100, e.g. 0-10 VAS score divided by 10 and multiplied by 100. 19 b SD imputed as variability measures not available 20 c SD estimated from p-value d SD estimated from standard error 21 e SD estimated from 95% Confidence Interval 22 f Sample size in the placebo group was divided by the number of groups to avoid double-counting 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 Cashin et al. 2021 Page 11 of 41 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 44 of 74

1 2 3 Supplementary file 7. Calculation of effect sizes for disability 4 5 Author, year Muscle Outcome Type of Type of Point estimate Mean (SD), converteda Number of participants 6 relaxant scale data measure (variability) extracted 7 medicine (range) extracted Muscle Comparator Muscle Comparator Muscle Comparator 8 Confidential: RelaxantFor ReviewRelaxant OnlyRelaxant 9 Immediate term (≤ 2 weeks) 10 Acute LBP 11 Friedman 2015 Cyclobenzaprine 0-24 Mean FV 8.2 8.9 34.2 (35)b 37.1 (34.8)b 108 107 12 RMDQ (95% CI) Friedman 2017 Diazepam 0-24 Mean CS -11 -11 -45.8 -45.8 (39.3)b 57 55 13 RMDQ (95% CI) (31.4)b 14 Friedman 2018 Orphenadrine 0-24 Mean CS -9.4 -10.9 -39.2 -45.4 (36.5)b 78 38g 15 RMDQ (95% CI) (37.)9b 16 Friedman 2018 Methocarbamol 0-24 Mean CS -8.1 -10.9 -33.8 -45.4 (36.5)b 80 38g 17 RMDQ (95% CI) (37.4)b 18 Friedman 2019 Baclofen 0-24 Mean CS -10.6 -11.1 -44.2 -46.3 (38.7)b 79 24g b 19 RMDQ (95% CI) (38.1) b g 20 Friedman 2019 Metaxalone 0-24 Mean CS -10.1 -11.1 -42.1 -46.3 (38.7) 76 25 RMDQ (95% CI) (39.2)b 21 Friedman 2019 Tizanidine 0-24 Mean CS -11.2 -11.1 -46.7 -46.3 (38.7)b 76 26g 22 RMDQ (95% CI) (36.5)b 23 Hindle 1972 Carisoprodol VRS-4 Mean FV 1.8 3.4 45 (30)c 85 (30)c 14 14 24 NCT00671879 Carisoprodol A 0-24 Mean (SE) CS -5 (0.6) -4.3 (0.7) -20.8 -17.9 (32.3)d 141 71g 25 2012 RMDQ (31.7)d 26 NCT00671879 Carisoprodol B 0-24 Mean (SE) CS -4.2 (0.6) -4.3 (0.7) -17.5 (31)d -17.9 (32.3)d 135 71g 27 2012 RMDQ e 28 Ralph 2008 Carisoprodol 0-24 Mean (p- FV 4.1 6.2 17.1 25.8 (37.2) 269 278 RMDQ value) (36.6)e 29 Serfer 2010 Carisoprodol A 0-24 Mean (SE) CS -5.7 (0.3) -4.4 (0.3) -23.8 -18.3 (21.7)d 269 133g 30 RMDQ (21.2)d 31 Serfer 2010 Carisoprodol B 0-24 Mean (SE) CS -5.4 (0.3) -4.4 (0.3) -22.5 -18.3 (21.7)d 259 132g 32 RMDQ (21.5)d 33 Mixed LBP 34 Aksoy 2002 Thiocolchicoside 0-24 Mean (SD) FV 7.2 (8.8) 11.8 (10) 30 (36.7) 49.2 (41.7) 174 155 35 RMDQ 36 Short term (3-13 weeks) Acute LBP 37 Friedman 2015 Cyclobenzaprine 0-24 Mean FV 4.5 3.8 18.8 15.8 (27.2)b 108 107 38 RMDQ (95% CI) (31.7)b 39 40 41 42 Cashin et al. 2021 Page 12 of 41 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 45 of 74 BMJ

1 2 3 Friedman 2017 Diazepam 0-24 Median FV 0 (0-1) 0 (0-6) 1.4 (3.2)f 8.3 (19.1)f 50 53 4 RMDQ (IQR) 5 Friedman 2018 Orphenadrine 0-24 Mean (SD) FV 5.6 (8) 3.8 (6.7) 23.3 (33.4) 16 (27.7) 69 34g 6 RMDQ g 7 Friedman 2018 Methocarbamol 0-24 Mean (SD) FV 4.9 (7.6) 3.8 (6.7) 20.6 (31.5) 16 (27.7) 70 34 RMDQ 8 Chronic LBP Confidential: For Review Only 9 Goforth 2015 Eszoplicone 0-24 Mean (SD) FV 6.6 (5.5) 7.9 (7) 27.5 (22.9) 33.1 (29.1) 32 20 10 RMDQ 11 Zaringhalam Baclofen A 0-24 Mean (SD) FV 8.8 (3.8) 9.8 (3.9) 36.7 (15.8) 40.8 (16.3) 20 20 12 2010 RMDQ 13 Zaringhalam Baclofen B 0-24 Mean (SD) FV 5.7 (1.4) 6.4 (2.9) 23.8 (5.8) 26.7 (12.1) 20 20 14 2010 RMDQ 15 16 SD, standard deviation; MD, mean difference; 95% CI, 95% confidence interval; FV, Final Value; CS, Change Score; RMDQ, Roland Morris Disability Questionnaire; VRS-4, Verbal Rating Scale 4 levels 17 a Mean and variability measures divided by the top number of scale and multiplied by 100, e.g. 0-24 RMDQ score divided by 24 and multiplied by 100. 18 b SD estimated from 95% Confidence Interval 19 c SD imputed as variability measures not available d SD estimated from standard error 20 e SD estimated from p-value 21 f SD estimated from median and IQR 22 g Sample size in the placebo group was divided by the number of groups to avoid double-counting 23

24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 Cashin et al. 2021 Page 13 of 41 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 46 of 74

1 2 3 Supplemental file 8. Characteristics of included studies 4 5 Study, Year Study sample Setting Number Test intervention, n Comparison Duration of Outcome Overall Source of 6 (Reference) of intervention, n treatment measure risk of data 7 Mean age (SD) and percentage relevant (Pain, Bias 8 femaleConfidential: (%) trial arms For Review OnlyDisability) 9 Akhter 20171 288 participants with mixed acute and India 2 Oral thiocolchicoside Oral diclofenac 7 days 10cm VAS, High Published 10 subacute LBP 150mg/day + sodium, 144 NA 11 diclofenac sodium, 12 Age and sex not reported 144 13 Aksoy 20022 329 participants with mixed acute and Turkey 2 Oral thiocolchicoside Standard treatment 5-7 days 100mm VAS, High Published 14 subacute LBP 16mg/day + standard (oral NSAID or RMDQ 15 treatment (NSAID or another analgesic), 16 thiocolchicoside group 39.7 (11) yrs, an analgesic), 174 155 67% female; standard treatment group 17 40.2 (11.3) yrs, 61% female 18 19 Aparna 20163 200 participants with acute LBP India 2 Oral thiocolchicoside Oral aceclofenac, 100 7 days 10cm VAS, High Published 20 8mg/day + NA 21 Age and sex not reported aceclofenac, 100 22 Baratta 19824 120 participants with acute LBP USA 2 Oral cyclobenzaprine Oral placebo, 60 10 days 10cm VAS, High Published 23 30mg/day, 60 NA 24 cyclobenzaprine group 35 yrs a, 41% 25 female; placebo group 38 yrs a, 41% 26 female 27 Berry (a) 19885 105 participants with acute LBP UK 2 Oral tizanidine Oral placebo + 7 days 100mm VAS, High Published 28 12mg/day + ibuprofen, ibuprofen, 54 NA 29 tizanidine group 43 (12.4) yrs, 47% 51 30 female; placebo group 42 (12.4) years, 43% female 31 32 Berry (b) 19886 112 participants with acute LBP UK 2 Oral tizanidine Oral placebo, 53 7 days 100mm VAS, High Published 33 12mg/day, 59 NA 34 tizanidine group 44 (13) yrs, 49% female; placebo group 38 (13) yrs, 35 49% female 36 37 Borenstein 40 participants with acute LBP USA 2 Oral cyclobenzaprine Oral naproxen, 20 14 days NR, VRS-4 b High Published 19907 30mg/day + naproxen, 38 20 39 40 41 42 Cashin et al. 2021 Page 14 of 41 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 47 of 74 BMJ

1 2 3 cyclobenzaprine group 37 yrs a, 35% 4 female; comparator group 37 yrs a, 5 25% female 6 Casale 19888 20 participants with acute LBP Italy 2 Oral dantrolene Oral placebo, 10 4 days NR, NR Moderate Published 7 25mg/day, 10 8 dantrolene group 46.7 yrs a, 30% Confidential:a For Review Only female; placebo group 47.1 yrs , 20% 9 female 10 11 Cogné 20179 19 participants with chronic LBP France 2 IM botulinum toxin A IM placebo, 10 Single dose 100mm VASb, High Published 12 200 units, 9 QBPDSb 13 (crossover) botulinum toxin A group 38.1 (5.94) yrs, 67% female; placebo group 38.2 14 (10.27) yrs, 100% female 15 16 Dapas 198510 200 participants with acute LBP USA 2 Oral baclofen range Oral placebo, 100 14 days VRS-5b, NA High Published 17 30-80mg/day, 100 baclofen group 42.7 yrs a, 48% female; 18 placebo group 41.8 yrs a, 56% female 19 20 Emrich 201511 202 participants with acute LBP Germany 2 Oral methocarbamol Oral placebo, 104 8 days 100mm VAS, High Published 21 4500mg/day, 98 NR methocarbamol group 45.3 (11) yrs, 22 63% female; placebo group 43.8 23 (11.6) yrs, 71% female 24 Fathie 196412 200 participants with acute LBP USA 2 Oral metaxalone Oral placebo, 99 7 days VRS-4b, NA High Published 25 3200mg/day, 101 26 Age and sex not reported 27 28 Foster 200113 31 participants with chronic LBP USA 2 IM botulinum toxin A IM placebo Single dose 10cm VASb, Low Published b 29 200 units ODI botulinum toxin A group 46.4 yrs a, 30 53% female; placebo group 47 yrs a, 31 50% female 32 Friedman 201514 323 participants with acute LBP USA 2 Oral cyclobenzaprine Oral placebo 10 days 10cm VAS, Low Published 33 range 5-30mg/day + +naproxen, 107 RMDQ 34 cyclobenzaprine group 38 (11) yrs, naproxen, 108 35 42% female; oxycodone group 39 (11) 36 yrs, 56% female [not synthesized]; placebo 39 (11) yrs, 50% female 37 38 39 40 41 42 Cashin et al. 2021 Page 15 of 41 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 48 of 74

1 2 3 Friedman 201715 114 participants with acute LBP USA 2 Oral diazepam range Oral placebo + 7 days VRS-4, RMDQ Low Published 4 5-20mg/day + naproxen 5 diazepam group 34 (12) yrs, 47% naproxen, 57 female; placebo group 38 (12) yrs, 6 42% female 7 8 Friedman 201816 240 participantsConfidential: with acute LBP USA 3 ForOral orphenadrine Review Oral placebo + Only7 days VRS-4, RMDQ Low Published 9 200mg/day + naproxen, 79 orphenadrine group 40 (12) yrs, 43% naproxen, 80 10 female; methocarbamol group 38 (12) 11 yrs, 51% female; placebo group 39 Oral methocarbamol 12 (12) yrs, 43% female range 2250- 4500mg/day + 13 naproxen, 81 14 15 Friedman 201917 320 participants with acute LBP USA 4 Oral tizanidine range Oral placebo + 7 days VRS-4, RMDQ Low Published 16 2-16mg/day + ibuprofen, 80 tizanidine group 40 (11) yrs, 48% ibuprofen, 80 17 female; metaxalone group 37 (10) yrs, 18 45% female; baclofen group 39 (12) Oral metaxalone 19 yrs, 29% female; placebo group 39 range 400- (11) yrs, 45% female 3200mg/day+ 20 ibuprofen, 80 21 22 Oral baclofen range 23 10-80mg/day + ibuprofen, 80 24 25 Goforth 201418 58 participants with chronic LBP USA 2 Oral eszopiclone Oral placebo + 28 days 100mm VAS, Low Published 26 3mg/day + naproxen, naproxen, 25 RMDQ 27 eszopiclone group 45.7 (11) yrs, 61% 33 female; placebo group 40.1 (12.8) yrs, 28 72% female 29 30 Gold 197819 60 participants with acute LBP USA 2 Oral orphenadrine Oral placebo, 20 7 days NR, NA High Published 31 200mg/day, 20 Age and sex not reported 32 33 Herskowitz 28 participants with subacute LBP USA 2 IM botulinum toxin A IM placebo, 15 Single dose 10cm VAS, High Published 34 200420 400 units, 13 NA (conference 35 Age and sex not reported abstract) 36 Hindle 197221 48 participants with acute LBP USA 2 Oral carisoprodol Oral placebo, 16 4 days 100mm VAS, High Published 37 1400mg/day, 16 VRS-4 38 39 40 41 42 Cashin et al. 2021 Page 16 of 41 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 49 of 74 BMJ

1 2 3 carisoprodol group 37 yrs a; 4 butabarbital group 34.6 yrs a; placebo a 5 group 43.5 yrs 6 Entire sample 44% female 7 8 Hingorani 196622 50 participantsConfidential: with acute LBP UK 2 ForIM diazepam 40mgReview + IM placebo + oral Only6 days NR, NA High Published 9 oral diazepam placebo, 25 Age not reported 8mg/day, 25 10 11 Entire sample 20% female 12 23 b 13 Jazayeri 2011 50 participants with chronic LBP Iran 2 IM botulinum toxin A IM placebo 25 Single dose 10cm VAS , High Published 200 units, 25 ODI b 14 botulinum toxin A group 41.7 yrs a, 15 52% female; placebo group 42.3 yrs a, 16 56% female 17 Ketenci 200524 97 participants with acute LBP Turkey 3 Oral thiocolchicoside Oral placebo, 27 7 days 10cm VAS, High Published 18 16mg/day, 38 NA 19 thiocolchicoside group 37 yrs a, 42% 20 female; tizanidine group 37 yrs a, 63% Oral tizanidine female; placebo group 40 yrs a, 52% 6mg/day, 32 21 female 22 23 Klinger 198825 80 participants with acute LBP USA 2 IV orphenadrine IV placebo, 40 Single dose VRS-4b, NA Low Published 24 60mg, 40 orphenadrine group 35.7 (12.4) yrs, 25 1% female; placebo group 31.9 (11.7) 26 yrs, 30% female 27 28 Lepisto 197926 30 participants with acute LBP Finland 2 Oral tizanidine Oral placebo, 15 7 days VRS-4, NA Moderate Published 6mg/day, 15 29 tizanidine group 42.5 yrs a, 47% 30 female; placebo group 40.8 yrs a, 53% 31 female 32 Machado 201627 43 participants with chronic LBP USA 2 IM botulinum toxin A IM placebo, 22 Single 10cm VASb, Moderate Published 33 range 500-1000 units, injection ODI b 34 botulinum toxin A group 51.3 yrs a, 21 35 67% female; placebo group 48.6 yrs a, 36 45% female 37 Pareek 200928 197 participants with acute LBP India 2 Oral tizanidine Oral aceclofenac, 96 7 days 10cm VAS, High Published 38 4mg/day + NA 39 aceclofenac, 101 40 41 42 Cashin et al. 2021 Page 17 of 41 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 50 of 74

1 2 3 tizanidine group 43.3 (12.7) yrs, 39% 4 female; comparator group 43.5 (10.9) 5 yrs, 40% female 6 Ralph 200829 562 participants with acute LBP USA 2 Oral carisoprodol Oral placebo, 285 7 days VRS-5, RMDQ High Published 7 1000mg/day, 277 8 carisoprodolConfidential: group 39.3 (11.8) yrs, For Review Only 47% female; comparator group 41.5 9 (11.7) yrs, 54% female 10 11 Salvini 198630 30 participants with LBP Italy 2 Oral dantrolene Oral ibuprofen, 15 8 days VRS-4b, NA High Published 12 1200mg/day + 13 Age and sex not reported ibuprofen, 15 14 Schliessbach 98 participants with chronic LBP Switzerlan 2 Oral clobazam 20mg, Oral placebo, 49 2 hours 11pt NRS, NA Low Published 15 201731 d 49 16 Age and sex not reported 17 (crossover) 18 Serfer 201032 828 participants with acute LBP USA 3 Oral carisoprodol Oral placebo, 276 7 days VRS-5, RMDQ High Published 19 (350mg) 1400mg/day, 20 carisoprodol (350mg) group 40.5 281 21 (12.4) yrs, 54% female; carisoprodol (250mg) group 40.9 (11.7) yrs, 51% Oral carisoprodol 22 female; placebo group 40.7 (13.1) yrs, (250mg) 1000mg/day, 23 59% female 271 24 Tervo 197633 50 participants with acute LBP Finland 2 IM orphenadrine 60mg IM placebo + oral 7-10 days NR, NR High Published 25 + oral orphenadrine paracetamol, 25 26 Age not reported 210mg/day & 27 paracetamol, 25 28 Entire sample 66% female 29 Thompson 76 participants with acute LBP UK 2 Oral tizanidine Oral placebo 10 days 100mm VASb, High Published 30 198334 6mg/day NA (conference 31 Age and sex not reported abstract) 32 33 Tüzün 200335 149 participants with acute LBP Turkey 2 IM thiocolchicoside IM placebo, 72 5 days 100mm VAS, High Published 34 8mg/day, 77 NA 35 thiocolchicoside group 40.7 (10.3) yrs, 36 48% female; placebo group 41 (11) 37 yrs, 56% female 38 Zaringhalam 84 participants with chronic LBP Iran 4 Oral baclofen No treatment, 21 35 days 100mm VAS, High Published 39 201036 30mg/day, 21 RMDQ 40 41 42 Cashin et al. 2021 Page 18 of 41 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 51 of 74 BMJ

1 2 3 baclofen group 55.1 (3.3) yrs; no Oral baclofen 4 treatment group 54.3 (4.2) yrs; 30mg/day + acupuncture group 54.2 (5.4) yrs; acupuncture, 21 Acupuncture, 21 5 baclofen + acupuncture group 54.2 6 (5.6) yrs 7 8 EntireConfidential: sample 0% female For Review Only 9 ACTRN1261600 Participants with acute LBP Australia 2 Oral zoplicone Oral placebo 14 days NA NA Clinical trial 10 001742637 7.5mg/day registry 11 12 (status: terminated) 13 14 EUCTR2017- 134 participants with acute LBP Greece 2 IM thiocolchicoside IM diclofenac Single NA NA Clinical trial 15 004530-2938 4mg + diclofenac injection registry 16 Age and sex not reported 17 EUCTR2019- Participants with acute LBP and/or Hungry 2 Oral tolperisone Oral placebo 14 days NA NA Clinical trial 18 001885-1439 sciatica registry 19 (status: ongoing) 20 IRCT201111090 46 participants with LBP Iran 2 Oral zolpidem Oral placebo 28 days NA NA Clinical trial 21 08035N440 5mg/day registry 22 Age and sex not reported 23 24 NCT0067187941 840 participants with acute LBP USA 3 Oral carisoprodol Oral placebo, 280 14 days 100mm VAS, High Clinical trial 25 (500mg) 1000mg/day, RMDQ registry carisoprodol (500mg) group 41.6 279 26 (11.8) yrs, 52% female; carisoprodol 27 (700mg) group 41.5 (12.4) yrs, 53% Oral carisoprodol 28 female; placebo group 41.4 (11.9) yrs, (700mg) 1400mg/day, 51% female 281 29 30 NCT0067150242 840 participants with acute LBP USA 3 Oral carisoprodol Oral placebo, 279 14 days 100mm VAS, High Clinical trial 31 (500mg) 1000mg/day, RMDQ b registry 32 carisoprodol (500mg) group 41.4 280 (12.6) yrs, 51% female; carisoprodol 33 (700mg) group 40.3 (13.1) yrs, 47% Oral carisoprodol 34 female; placebo group 40.9 (12.7) yrs, (700mg) 1400mg/day, 35 49% female 281 36 NCT0081798643 161 participants with acute LBP USA 4 Oral arbaclofen Oral placebo 14 days NA NA Clinical trial 37 placarbil (20mg) registry 38 Age and sex not reported 40mg/day 39 40 41 42 Cashin et al. 2021 Page 19 of 41 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 52 of 74

1 2 3 Oral arbaclofen 4 placarbil (30mg) 5 60mg/day 6 Oral arbaclofen 7 placarbil (40mg) 8 Confidential: For80mg/day Review Only 9 NCT0040441744 Participants with chronic LBP USA 4 IM botulinum toxin A IM placebo Single dose NA NA Clinical trial 10 registry 11 (crossover, 12 status: active not recruiting) 13 14 NCT0038457945 Participants with acute LBP USA 2 IM botulinum toxin B IM placebo Single dose NA NA Clinical trial 15 registry 16 (status: terminated) 17 18 NCT0038437146 Participants with subacute LBP USA 2 IM botulinum toxin A IM placebo Single dose NA NA Clinical trial 19 registry 20 (status: terminated) 21 22 NCT0288753447 Participants with acute LBP Not 5 Oral tizanidine Oral placebo Not reported NA NA Clinical trial 23 reported registry 24 (status: Oral SPARC1401-low withdrawn) dose 25 26 Oral SPARC1401-mid 27 dose 28 Oral SPARC1401-high 29 dose 30 31 NCT0158750848 Participants with acute LBP Brazil 3 Oral cyclobenzaprine Oral meloxicam & 7 days NA NA Clinical trial 32 20mg/day cyclobenzaprine registry (status: 33 withdrawn) Oral meloxicam 34 35 a Standard deviation not reported. b Data not available. Abbreviations: LBP, Low Back Pain; SD, Standard Deviation; IM, Intramuscular; IV, Intravenous; NA, Not Applicable; NR Not 36 Reported; NRS, Numerical Rating Scale; VAS, Visual Rating Scale; VRS-4, Verbal Rating Scale 4 levels; VRS-5, Verbal Rating Scale 5 levels; RMDQ, Roland Morris Disability 37 38 Questionnaire; QBPDS, Quebec Back Pain Disability Scale 39 40 41 42 Cashin et al. 2021 Page 20 of 41 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 53 of 74 BMJ

1 2 3 References: 4 5 1. Akhter N, Siddiq MZ. Comparative efficacy of diclofenac sodium alone and in combination with thiocolchicoside in patients with low back 6 pain. Med Forum. 2017;28(11):93-96. 7 2. Aksoy C, Karan A, Diraçoǧlu D. Low back pain: Results of an open clinical trial comparing the standard treatment alone to the 8 Confidential: For Review Only 9 combination of standard treatment and thiocolchicoside. J Orthop Traumatol. 2002;3(2):103-108. doi:10.1007/s101950200036 10 3. Aparna P, Geetha P, Shanmugasundaram P. Comparison of aceclofenac and combination (Aceclofenac + thiocolchicoside) therapy in 11 acute low back pain patients. Res J Pharm Technol. 2016;9(11):1927-1929. doi:10.5958/0974-360X.2016.00394.2 12 4. Baratta RR. A double-blind study of cyclobenzaprine and placebo in the treatment of acute musculoskeletal conditions of the low back. 13 Curr Ther Res. 1982;32(5):646-652. 14 5. Berry H, Hutchinson DR. A Multicentre Placebo-Controlled Study in General Practice to Evaluate the Efficacy and Safety of Tizanidine in 15 Acute Low-Back Pain. J Int Med Res. 1988;16(2):75-82. doi:10.1177/030006058801600201 16 6. Berry H, Hutchinson DR. Tizanidine and Ibuprofen in Acute Low-Back Pain: Results of a Double-Blind Multicentre Study in General 17 Practice. J Int Med Res. 1988;16(2):83-91. doi:10.1177/030006058801600202 18 7. Borenstein DG, Lacks S, Wiesel SW. Cyclobenzaprine and naproxen versus naproxen alone in the treatment of acute low back pain and 19 muscle spasm. Clin Ther. 1990;12(2):125-131. 20 8. Casale R. Acute low back pain: Symptomatic treatment with a muscle relaxant drug. Clin J Pain. 1988;4:81-88. 21 9. Cogné M, Petit H, Creuzé A, Liguoro D, de Seze M. Are paraspinous intramuscular injections of botulinum toxin a (BoNT-A) efficient in 22 the treatment of chronic low-back pain? A randomised, double-blinded crossover trial. BMC Musculoskelet Disord. 2017;18(1):454. 23 doi:10.1186/s12891-017-1816-6 24 10. Dapas F, Hartman SF, Martinez L, et al. Baclofen for the treatment of acute low-back syndrome: A double-blind comparison with 25 placebo. Spine (Phila Pa 1976). 1985;10(4):345-349. doi:10.1097/00007632-198505000-00010 26 11. Emrich OMD, Milachowski KA, Strohmeier M. Methocarbamol bei akuten Rückenschmerzen: Eine randomisierte, doppelblinde, 27 28 placebokontrollierte Studie. MMW-Fortschritte der Medizin. 2015;157:9-16. doi:10.1007/s15006-015-3307-x 29 12. Fathie K. A second look at skeletal muscle relaxant: a double-blind study with metaxalone. Curr Ther Res. 1964;6(11):677-683. 30 13. Foster L, Clapp L, Erickson M, Jabbari B. Botulinum toxin A and chronic low back pain a randomized, double-blind study. Neurology. 31 2001;56(10):1290-1293. doi:10.1212/WNL.56.10.1290 32 14. Friedman BW, Dym AA, Davitt M, et al. Naproxen with cyclobenzaprine, oxycodone/acetaminophen, or placebo for treating acute low 33 back pain: A randomized clinical trial. JAMA - J Am Med Assoc. 2015;314(15):1572-1580. doi:10.1001/jama.2015.13043 34 15. Friedman BW, Irizarry E, Solorzano C, et al. Diazepam Is No Better Than Placebo When Added to Naproxen for Acute Low Back Pain. 35 Ann Emerg Med. 2017;70(2):169-176.e1. doi:10.1016/j.annemergmed.2016.10.002 36 16. Friedman BW, Cisewski D, Irizarry E, et al. A Randomized, Double-Blind, Placebo-Controlled Trial of Naproxen With or Without 37 Orphenadrine or Methocarbamol for Acute Low Back Pain. Ann Emerg Med. 2018;71(3):348-356.e5. 38 doi:10.1016/j.annemergmed.2017.09.031 39 17. Friedman BW, Irizarry E, Solorzano C, et al. A Randomized, Placebo-Controlled Trial of Ibuprofen Plus Metaxalone, Tizanidine, or 40 41 42 Cashin et al. 2021 Page 21 of 41 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 54 of 74

1 2 3 Baclofen for Acute Low Back Pain. Ann Emerg Med. 2019;74(4):512-520. doi:10.1016/j.annemergmed.2019.02.017 4 18. Goforth HW, Preud’homme XA, Krystal AD. A Randomized, Double-Blind, Placebo-Controlled Trial of Eszopiclone for the Treatment of 5 Insomnia in Patients with Chronic Low Back Pain. Sleep. 2014;37(6):1053-1060. doi:10.5665/sleep.3760 6 19. Gold RH. Orphenadrine citrate: sedative or muscle relaxant? Clin Ther. 1978;1(6):451-453. 7 20. Herskowitz A. BOTOX (Botulinum Toxin Type A) treatment of patients with sub-acute low back pain: A randomized, double blind, 8 placebo-controlledConfidential: study. J Pain. 2004;5(3):S62. doi:10.1016/j.jpain.2004.02.214 For Review Only 9 10 21. Hindle T, Palma L. Comparison of carisoprodol, butabarbital, and placebo in treatment of low back syndrome. Calif Med. 1972;117:7-11. 11 22. Hingorani K. Diazepam in backache: A double-blind controlled trial. Ann Phys Med. 1966;8(8):303-306. 12 23. Jazayeri SM, Ashraf A, Fini HM, Karimian H, Nasab M V. Efficacy of Botulinum Toxin Type A for Treating Chronic Low Back Pain. 13 Anesthesiol Pain Med. 2011;1(2):77-80. doi:10.5812/kowsar.22287523.1845 14 24. Ketenci A, Ozcan E, Karamursel S. Assessment of efficacy and psychomotor performances of thiocolchicoside and tizanidine in patients 15 with acute low back pain. Int J Clin Pract. 2005;59(7):764-770. doi:10.1111/j.1742-1241.2004.00454.x 16 25. Klinger N., Wilson R., Kanniainen C., Wagenknecht K., Re O., Gold R. Intravenous oprhenadrine for the treatment of lumbar 17 paravertebral muscle strain. Curr Ther Res. 1988;43(2):247-254. 18 26. Lepisto P. A Comparative Trial of DS 103-282 and Placebo in the Treatment of Acute Skeletal Muscle Spasms Due to Disorders of the 19 Back. Curr Ther Res. 1979;26(4):454-459. 20 27. Machado D, Kumar A, Jabbari B. Abobotulinum toxin A in the treatment of chronic low back pain. Toxins (Basel). 2016;8(12). 21 doi:10.3390/toxins8120374 22 28. Pareek A, Chandurkar N, Chandanwale AS, Ambade R, Gupta A, Bartakke G. Aceclofenac-tizanidine in the treatment of acute low back 23 pain: A double-blind, double-dummy, randomized, multicentric, comparative study against aceclofenac alone. Eur Spine J. 24 2009;18(12):1836-1842. doi:10.1007/s00586-009-1019-4 25 29. Ralph L, Look M, Wheeler W, Sacks H. Double-blind, placebo-controlled trial of carisoprodol 250-mg tablets in the treatment of acute 26 lower-back spasm. Curr Med Res Opin. 2008;24(2):551-558. doi:10.1185/030079908X261014 27 30. Salvini S, Antonelli S, De Micheli G, Marchetti M. Dantrolene sodium in low back pain and cervico brachialgia treatment: a controlled 28 29 study. Curr Ther Res. 1986;39(2):172-177. 30 31. Schliessbach J, Vuilleumier PH, Siegenthaler A, et al. Analgesic effect of clobazam in chronic low-back pain but not in experimentally 31 induced pain. Eur J Pain (United Kingdom). 2017;21(8):1336-1345. doi:10.1002/ejp.1032 32 32. Serfer GT, Wheeler WJ, Sacks HJ. Randomized, double-blind trial of carisoprodol 250 mg compared with placebo and carisoprodol 350 33 mg for the treatment of low back spasm. Curr Med Res Opin. 2010;26(1):91-99. doi:10.1185/03007990903382428 34 33. Tervo T, Petaja L, Lepisto P. A controlled clinical trial of a muscle relaxant analgesic combination in the treatment of acute lumbago. Br J 35 Clin Pract. 1976;30(3):62-64. 36 34. Thompson M, Kennedy G. Treatment of acute low back pain: compartive trial of two muscle relaxants, tizanidine and chlormezanone 37 with placebo. In: Scandinavian Journal of Rheumatology. Vol 12. ; 1983:4-40. doi:10.3109/03009748309118006 38 35. Tüzün F, Ünalan H, Öner N, et al. Multicenter, randomized, double-blinded, placebo-controlled trial of thiocolchicoside in acute low back 39 pain. Jt Bone Spine. 2003;70(5):356-361. doi:10.1016/S1297-319X(03)00075-7 40 41 42 Cashin et al. 2021 Page 22 of 41 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 55 of 74 BMJ

1 2 3 36. Zaringhalam J, Manaheji H, Rastqar A, Zaringhalam M. Reduction of chronic non-specific low back pain: A randomised controlled clinical 4 trial on acupuncture and baclofen. Chin Med. 2010;5:1-7. doi:10.1186/1749-8546-5-15 5 37. ACTRN12616000017426. A randomised controlled feasibility study of managing sleep with Zopiclone in participants with acute low back 6 pain and sleep disturbances. Australian New Zealand Clinical Trials Registry. 7 38. EUCTR2017-004530-29. No Title. Clinicaltrialsregister.eu. https://www.clinicaltrialsregister.eu/ctr-search/trial/2017-004530-29/HU. 8 Published 2017.Confidential: For Review Only 9 10 39. EUCTR2019-001885-14. No Title. Clinicaltrialsregister.eu. https://www.clinicaltrialsregister.eu/ctr-search/trial/2019-001885-14/HU. 11 Published 2019. 12 40. IRCT20111109008035N4. The study of efficacy of melatonin, and zolpidem on the sleep quality, and severity of pain in the patients with 13 chronic non-specific low back pain. ISRCTN. 14 41. NCT00671879. Study to Evaluate Two Formulations of Carisoprodol in Subjects With Musculoskeletal Spasm of the Lower Back. 15 ClinicalTrials.gov. https://clinicaltrials.gov/ct2/show/NCT00671879. Published 2008. 16 42. NCT00671502. A Study to Evaluate Two Formulations of Carisoprodol in Subjects With Musculoskeletal Spasm of the Lower Back. 17 ClinicalTrials.gov. 18 43. NCT00817986. A Study to Evaluate the Safety and Tolerability of Arbaclofen Placarbil (XP19986) in Subjects With Acute Back Spasms. 19 ClinicalTrials.gov. 20 44. NCT00404417. Botulinum Toxin A for the Treatment of Chronic Lumbar Back Pain. ClinicalTrials.gov. 21 45. NCT00384579. Pilot Study to Assess the Efficacy of Botulinum Toxin B on Pain and Disability in Subjects With Acute Low Back Pain. 22 ClinicalTrials.gov. 23 46. NCT00384371. Pilot Study to Assess the Efficacy of Botulinum Toxin A Treatments on Pain and Disability in Sub-Acute Low Back Pain. 24 ClinicalTrials.gov. 25 47. NCT02887534. Evaluation of Efficacy and Safety of SPARC1401 in Acute Low Back Pain. ClinicalTrials.gov. 26 48. NCT01587508. Study Comparing A New Drug Containing The Combination Meloxicam And Cyclobenzaprine In The Treatment Of Acute 27 Lumbago. ClinicalTrials.gov. 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 Cashin et al. 2021 Page 23 of 41 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 56 of 74

1 2 3 Supplemental file 9. Narrative description of trials not included in meta-analysis for 4 5 pain intensity 6 7 Study, Year (Reference) Outcome (Pain intensity) 8 1 9 Borenstein 1990 “The total pain scores, as determined by the patients daily 10 and physicians during scheduled visits, were not 11 significantly different.” 12 Confidential: For Review Only 13 Casale 19882 “VAS [visual analogue scale] pain measurements during the 14 maximal voluntary movements showed a decrease in pain 15 rating clearly in favor of dantrolene, with a percentage 16 variation of 50% for the drug and 8.6% for placebo. 17 Statistical comparison between the two treatments showed 18 dantrolene to have a higher effectiveness (p<0.001).” 19 20 Cogné 20173 First phase crossover data was not available. The study 21 found “no significant difference between the groups’ 22 (crossover) [botulinum toxin A vs placebo] average LBP [low back pain] 23 24 during the last 8 days at Day 30 (p = 0.97)”.

25 4 26 Dapas 1985 Patients were categorised into subgroups based on low 27 back symptom severity, moderate initial pain and sever or 28 extremely severe initial pain. “When the severity of 29 symptoms at visits 2 and 3 [day 4 and 10] was compared 30 with baseline values at visit 1 [day 1] within the placebo and 31 the baclofen treatment groups, all efficacy variables 32 [including local pain in lumbar area] showed a statistically 33 significant (P<0.05) improvement for the severe- and 34 moderate-pain groups.” 35 36 Emrich 20155 “The proportion of patients treated with methocarbamol who 37 achieved a pain-free state rose more rapidly to over 80% 38 39 and accordingly the proportion of patients who were not yet 40 pain-free after 8 days is below 20% - in contrast to ~ 60% in 41 the placebo group” 42 43 Fathie 19646 “A medically significant response was observed in 69.6% of 44 the 46 metaxalone-treated patients who complete the 45 course of therapy and returned for re-examination”. 46 Compared to “17.4% of the placebo-treatment patients who 47 completed the course of therapy [and] showed a medically 48 significant improvement”. 49 50 Foster 20017 “At 3 weeks, 11 of 15 patients who received botulinum toxin 51 (73.3%) had >50% pain relief vs four of 16 (25%) in the 52 saline group (p < 0.012). At 8 weeks, nine of 15 (60%) in 53 54 the botulinum toxin group and two of 16 (12.5%) in the 55 saline group had relief (p < 0.009).” 56 8 57 Gold 1978 At the 48-hour evaluation, 7/20 patients treated with 58 orphenadrine improved compared to 0/20 in the placebo 59 group. 60

Cashin et al. 2021 Page 24 of 41 https://mc.manuscriptcentral.com/bmj Page 57 of 74 BMJ

1 2 3 Hingorani 19669 “Of the 25 patients in the placebo group, 18 showed 4 improvement, 5 showed no change, and 2 were worse. Of 5 the 25 patients in the diazepam group, 19 showed 6 improvement, 5 showed no change, and 1 was worse. The 7 8 difference would therefore seem to be marginal, patients in 9 the treated group having almost no better results than those 10 in the placebo group.” 11 12 JazayeriConfidential: 201110 “After 4 weeks, For 76% ofReview patients in the BoNT Only-A [botulinum 13 toxin A] group reported pain relief compared to 20% in the 14 saline group (P < 0. 005). Additionally, greater pain relief 15 was experienced by patients in the BoNT-A group at 8 16 weeks (64% vs. 12%; P < 0. 001).” 17 18 Klinger 198811 “Based on both the physicians’ evaluations of signs and 19 symptoms and the patients’ assessments of pain, 20 intravenous orphenadrine was highly effective compared 21 with placebo in reducing these patients’ lumbar 22 23 paravertebral muscle pain and spasm.”

24 12 25 Machado 2016 “The primary outcome of this study was the proportion of 26 responders with a visual analogue scale (VAS) of <4 at 6 27 weeks. At 6 weeks, 5 subjects in the [abobotulinum toxin A] 28 toxin group and 3 subjects in the placebo group (28% and 29 16%) met this criterion (p = 0.4470).” 30 31 Salvini 198613 There was no significant difference between the groups 32 dantrolene and ibuprofen vs ibuprofen for pain on 33 movement and pain at rest at 4 and 8 days of treatment. 34 35 Schliessbach 201714 First phase crossover data was not available. The study 36 found “pain intensity in the supine position was significantly 37 (crossover) reduced by clobazam compared to active placebo (60 min: 38 2.9 vs. 3.5, p = 0.008; 90 min: 2.7 vs. 3.3, p = 0.024; 120 39 min: 2.4 vs. 3.1, p = 0.005). Pain intensity in the sitting 40 41 position was not significantly different between groups.” 42 15 43 Tervo 1976 No statistically significant difference was observed for 44 symptom relief from low back for orphenadrine vs saline 45 immediately after the injection or at 7-10 days follow-up. 46 47 Thompson 198316 Tizanidine was “generally better than placebo and 48 significantly so in respect of VAS [visual analogue scale 49 pain intensity]”. 50 51 ACTRN1261600001742617 Trial terminated 52 53 54 55 EUCTR2017-004530-2918 No data available 56 57 EUCTR2019-001885-1419 Trial ongoing 58 59 60

Cashin et al. 2021 Page 25 of 41 https://mc.manuscriptcentral.com/bmj BMJ Page 58 of 74

1 2 3 IRCT20111109008035N420 No data available 4 5 NCT0081798621 No data available 6 7 8 9 NCT0040441722 Trial active but not recruiting 10 11 12 Confidential: For Review Only 13 NCT0038457923 Trial terminated 14 15 16 17 NCT0288753424 Trial withdrawn 18 19 20 21 NCT0158750825 Trial withdrawn 22 23 24 25 26 27 References: 28 29 1. Borenstein DG, Lacks S, Wiesel SW. Cyclobenzaprine and naproxen versus 30 naproxen alone in the treatment of acute low back pain and muscle spasm. Clin Ther. 31 1990;12(2):125-131. 32 2. Casale R. Acute low back pain: Symptomatic treatment with a muscle relaxant drug. 33 Clin J Pain. 1988;4:81-88. 34 3. Cogné M, Petit H, Creuzé A, Liguoro D, de Seze M. Are paraspinous intramuscular 35 injections of botulinum toxin a (BoNT-A) efficient in the treatment of chronic low-back 36 pain? A randomised, double-blinded crossover trial. BMC Musculoskelet Disord. 37 2017;18(1):454. doi:10.1186/s12891-017-1816-6 38 39 4. Dapas F, Hartman SF, Martinez L, et al. Baclofen for the treatment of acute low-back 40 syndrome: A double-blind comparison with placebo. Spine (Phila Pa 1976). 41 1985;10(4):345-349. doi:10.1097/00007632-198505000-00010 42 5. Emrich OMD, Milachowski KA, Strohmeier M. Methocarbamol bei akuten 43 Rückenschmerzen: Eine randomisierte, doppelblinde, placebokontrollierte Studie. 44 MMW-Fortschritte der Medizin. 2015;157:9-16. doi:10.1007/s15006-015-3307-x 45 6. Fathie K. A second look at skeletal muscle relaxant: a double-blind study with 46 metaxalone. Curr Ther Res. 1964;6(11):677-683. 47 7. Foster L, Clapp L, Erickson M, Jabbari B. Botulinum toxin A and chronic low back pain 48 a randomized, double-blind study. Neurology. 2001;56(10):1290-1293. 49 doi:10.1212/WNL.56.10.1290 50 8. Gold RH. Orphenadrine citrate: sedative or muscle relaxant? Clin Ther. 51 1978;1(6):451-453. 52 9. Hingorani K. Diazepam in backache: A double-blind controlled trial. Ann Phys Med. 53 1966;8(8):303-306. 54 10. Jazayeri SM, Ashraf A, Fini HM, Karimian H, Nasab M V. Efficacy of Botulinum Toxin 55 Type A for Treating Chronic Low Back Pain. Anesthesiol Pain Med. 2011;1(2):77-80. 56 doi:10.5812/kowsar.22287523.1845 57 58 11. Klinger N., Wilson R., Kanniainen C., Wagenknecht K., Re O., Gold R. Intravenous 59 oprhenadrine for the treatment of lumbar paravertebral muscle strain. Curr Ther Res. 60 1988;43(2):247-254.

Cashin et al. 2021 Page 26 of 41 https://mc.manuscriptcentral.com/bmj Page 59 of 74 BMJ

1 2 3 12. Machado D, Kumar A, Jabbari B. Abobotulinum toxin A in the treatment of chronic low 4 back pain. Toxins (Basel). 2016;8(12). doi:10.3390/toxins8120374 5 13. Salvini S, Antonelli S, De Micheli G, Marchetti M. Dantrolene sodium in low back pain 6 and cervico brachialgia treatment: a controlled study. Curr Ther Res. 1986;39(2):172- 7 177. 8 14. Schliessbach J, Vuilleumier PH, Siegenthaler A, et al. Analgesic effect of clobazam in 9 10 chronic low-back pain but not in experimentally induced pain. Eur J Pain (United 11 Kingdom). 2017;21(8):1336-1345. doi:10.1002/ejp.1032 12 15. TervoConfidential: T, Petaja L, Lepisto P. A controlled For clinicalReview trial of a muscle Only relaxant analgesic 13 combination in the treatment of acute lumbago. Br J Clin Pract. 1976;30(3):62-64. 14 16. Thompson M, Kennedy G. Treatment of acute low back pain: compartive trial of two 15 muscle relaxants, tizanidine and chlormezanone with placebo. In: Scandinavian 16 Journal of Rheumatology. Vol 12. ; 1983:4-40. doi:10.3109/03009748309118006 17 17. ACTRN12616000017426. A randomised controlled feasibility study of managing sleep 18 with Zopiclone in participants with acute low back pain and sleep disturbances. 19 Australian New Zealand Clinical Trials Registry. 20 18. EUCTR2017-004530-29. No Title. Clinicaltrialsregister.eu. 21 https://www.clinicaltrialsregister.eu/ctr-search/trial/2017-004530-29/HU. Published 22 2017. 23 19. EUCTR2019-001885-14. No Title. Clinicaltrialsregister.eu. 24 https://www.clinicaltrialsregister.eu/ctr-search/trial/2019-001885-14/HU. Published 25 2019. 26 20. IRCT20111109008035N4. The study of efficacy of melatonin, and zolpidem on the 27 sleep quality, and severity of pain in the patients with chronic non-specific low back 28 29 pain. ISRCTN. 30 21. NCT00817986. A Study to Evaluate the Safety and Tolerability of Arbaclofen Placarbil 31 (XP19986) in Subjects With Acute Back Spasms. ClinicalTrials.gov. 32 22. NCT00404417. Botulinum Toxin A for the Treatment of Chronic Lumbar Back Pain. 33 ClinicalTrials.gov. 34 23. NCT00384579. Pilot Study to Assess the Efficacy of Botulinum Toxin B on Pain and 35 Disability in Subjects With Acute Low Back Pain. ClinicalTrials.gov. 36 24. NCT02887534. Evaluation of Efficacy and Safety of SPARC1401 in Acute Low Back 37 Pain. ClinicalTrials.gov. 38 25. NCT01587508. Study Comparing A New Drug Containing The Combination 39 Meloxicam And Cyclobenzaprine In The Treatment Of Acute Lumbago. 40 ClinicalTrials.gov. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 27 of 41 https://mc.manuscriptcentral.com/bmj BMJ Page 60 of 74

1 2 3 Supplemental file 10. Risk of bias assessments 4

5

6

7

treat treat 8 Confidential: For Review- Only to

9 - sequence sequence assessment

10 providers) - Blinding reporting analysis? 11 interventions Allocation Drop Outs Other bias generation assessors) - Compliance 12 concealment (Care Co Intention Selective outcome Blinding (Patients) Random Blinding (Outcome

13 Overall Risk of Bias Similarity at baseline

Study Year Timing of 14 Low Low Low Low Low High Low Low 15 Unclear Unclear Unclear Unclear Unclear High Fathie 1964 risk risk risk risk risk risk risk risk 16 Low Low Low Low Low High Low 17 Unclear Unclear Unclear Unclear Unclear Unclear High Hingorani 1966 risk risk risk risk risk risk risk 18 Low Low Low Low Low High Low Low Low Unclear Unclear Unclear Unclear High 19 Hindle 1972 risk risk risk risk risk risk risk risk risk 20 Low Low Low Low High Low Low Unclear Unclear Unclear Unclear Unclear Unclear High 21 Tervo 1976 risk risk risk risk risk risk risk 22 Low Low Low Low Unclear Unclear Unclear Unclear Unclear Unclear Unclear Unclear Unclear High 23 Gold 1978 risk risk risk risk 24 Low Low Low Low Low Low Low Low Low Unclear Unclear Unclear Unclear Moderate 25 Lepisto 1979 risk risk risk risk risk risk risk risk risk 26 Low Low Low Low Low High Low Low Low Low Unclear Unclear Unclear High 27 Baratta 1982 risk risk risk risk risk risk risk risk risk risk 28 Low Low Low Low Unclear Unclear Unclear Unclear Unclear Unclear Unclear Unclear Unclear High 29 Thompson 1983 risk risk risk risk 30 Low Low Low High High High Low Low Unclear Unclear Unclear Unclear Unclear High 31 Dapas 1985 risk risk risk risk risk risk risk risk 32 High High High Low Low Unclear Unclear Unclear Unclear Unclear Unclear Unclear Unclear High 33 Salvini 1986 risk risk risk risk risk 34 Low Low Low Low High Low Low Low Unclear Unclear Unclear Unclear Unclear High 35 Berry (a) 1988 risk risk risk risk risk risk risk risk 36 Low Low Low Low Low Low Low Low Unclear Unclear Unclear Unclear Unclear High 37 Berry (b) 1988 risk risk risk risk risk risk risk risk 38 Low Low Low Low Low Low Low Low Low Unclear Unclear Unclear Unclear Moderate 39 Casale 1988 risk risk risk risk risk risk risk risk risk 40 41 42 Cashin et al. 2021 Page 28 of 41 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 61 of 74 BMJ

1 2 3 Low Low Low Low Low Low Low Low Low Low Unclear Unclear Unclear Low 4 Klinger 1988 risk risk risk risk risk risk risk risk risk risk 5 High High High Low Low Low Low Low Low Unclear Unclear Unclear Unclear High 6 Borenstein 1990 risk risk risk risk risk risk risk risk risk 7 Low Low Low Low Low Low Low Low Low Low Low Low Unclear Low 8 Foster 2001Confidential: risk risk risk risk risk For risk Reviewrisk risk risk Onlyrisk risk risk 9 Low High High High Low Low Low Low Low Unclear Unclear Unclear Unclear High 10 Aksoy 2002 risk risk risk risk risk risk risk risk risk 11 Low Low Low Low Low Low Low Low Unclear Unclear Unclear Unclear Unclear High 12 Tuzun 2003 risk risk risk risk risk risk risk risk 13 Low Low Low Low Low Low Unclear Unclear Unclear Unclear Unclear Unclear Unclear High 14 Herskowitz 2004 risk risk risk risk risk risk 15 Low Low Low Low Low Low Low High Low Unclear Unclear Unclear Unclear High 16 Ketenci 2005 risk risk risk risk risk risk risk risk risk Low Low Low Low Low Low Low Low Low Low High 17 Unclear Unclear High 18 Ralph 2008 risk risk risk risk risk risk risk risk risk risk risk Low High Low Low Low High 19 Unclear Unclear Unclear Unclear Unclear Unclear Unclear High 20 Pareek 2009 risk risk risk risk risk risk Low Low Low Low Low Low Low Low Low High Low High 21 Unclear High 22 Serfer 2010 risk risk risk risk risk risk risk risk risk risk risk risk Low High High High High Low High Low Low Low Low Low 23 Unclear High 24 Zaringhalam 2010 risk risk risk risk risk risk risk risk risk risk risk risk Low High Low Low Low Low Low Low Low Low 25 Unclear Unclear Unclear High 26 Jazayeri 2011 risk risk risk risk risk risk risk risk risk risk Low Low Low Low High Low 27 Unclear Unclear Unclear Unclear Unclear Unclear Unclear High 28 NCT00671502 2011 risk risk risk risk risk risk Low Low Low Low Low Low 29 Unclear Unclear Unclear Unclear Unclear Unclear Unclear High 30 NCT00671879 2012 risk risk risk risk risk risk Low Low Low Low Low Low Low Low Low Low 31 Unclear Unclear Unclear Low 32 Goforth 2014 risk risk risk risk risk risk risk risk risk risk Low Low Low High Low Low Low Low Low 33 Unclear Unclear Unclear Unclear High Emrich 2015 risk risk risk risk risk risk risk risk risk 34 Low Low Low Low Low Low Low Low Low Low 35 Unclear Unclear Unclear Low Friedman 2015 risk risk risk risk risk risk risk risk risk risk 36 High High High High Low Low 37 Unclear Unclear Unclear Unclear Unclear Unclear Unclear High Aparna 2016 risk risk risk risk risk risk 38 Low Low Low Low Low Low Low Low Low 39 Unclear Unclear Unclear Unclear Moderate Machado 2016 risk risk risk risk risk risk risk risk risk 40 41 42 Cashin et al. 2021 Page 29 of 41 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 62 of 74

1 2 3 High High High Low Low Unclear Unclear Unclear Unclear Unclear Unclear Unclear Unclear High 4 Akhter 2017 risk risk risk risk risk 5 Low Low Low Low Low High Low Low Low Low Low Low Unclear High 6 Cogne 2017 risk risk risk risk risk risk risk risk risk risk risk risk 7 Low Low Low Low Low Low Low Low Low Low Unclear Unclear Unclear Low 8 Friedman 2017Confidential: risk risk risk risk For Reviewrisk risk risk Onlyrisk risk risk 9 Low Low Low Low Low Low Low Low Low Low Low Unclear Unclear Low 10 Schliessbach 2017 risk risk risk risk risk risk risk risk risk risk risk 11 Low Low Low Low Low Low Low Low Low Low Low Unclear Unclear Low 12 Friedman 2018 risk risk risk risk risk risk risk risk risk risk risk 13 Low Low Low Low Low Low Low Low Low Low Unclear Unclear Unclear Low 14 Friedman 2019 risk risk risk risk risk risk risk risk risk risk 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 Cashin et al. 2021 Page 30 of 41 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 63 of 74 BMJ

1 2 3 Supplemental file 11. Forest plot pain intensity 3-13 weeks 4 5 6 Acute LBP – Non-benzodiazepine antispasmodic 7 8 9 10 11 12 Confidential: For Review Only 13 14 15 16 17 18 19 20 Acute LBP – Antispastic 21 22 23 24 25 26 27

28 29 Acute LBP – Benzodiazepine 30 31 32 33 34 35 36 37 38 39 Subacute LBP – Miscellaneous 40 41 42 43 44 45 46

47 48 Chronic LBP – Antispastic 49 50 51 52 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 31 of 41 https://mc.manuscriptcentral.com/bmj BMJ Page 64 of 74

1 2 3 Chronic LBP – Miscellaneous 4 5 6 7 8 9 10 11 12 Confidential: For Review Only 13 Mixed LBP – Non-benzodiazepine antispasmodic 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 32 of 41 https://mc.manuscriptcentral.com/bmj Page 65 of 74 BMJ

1 2 3 Supplemental file 12. Forest plot disability ≤ 2 weeks 4 5 Acute LBP – Non-benzodiazepine antispasmodic 6 7 8 9 10 11 12 Confidential: For Review Only 13 14 15 16 17 18 19 20 21 22 23 Acute LBP – Antispastic 24 25 26 27 28 29 30 31 32 Acute LBP – Benzodiazepine 33 34 35 36 37 38 39 40 41 Mixed LBP – Non-benzodiazepine antispasmodic 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 33 of 41 https://mc.manuscriptcentral.com/bmj BMJ Page 66 of 74

1 2 3 Supplemental file 13. Forest plot disability 3-13 weeks 4 5 Acute LBP – Non-benzodiazepine antispasmodic 6 7 8 9 10 11 12 Confidential: For Review Only 13 14 15 16 Acute LBP – Benzodiazepine 17 18 19 20 21 22 23 24 25 26 Chronic LBP – Antispastic 27 28 29 30 31 32 33 34 35 36 37 Chronic LBP – Miscellaneous 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 34 of 41 https://mc.manuscriptcentral.com/bmj Page 67 of 74 BMJ

1 2 3 Supplemental file 14. Forest plot acceptability 4 5 Acute LBP – Non-benzodiazepine antispasmodic 6 7 8 9 10 11 12 Confidential: For Review Only 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 Chronic LBP – Antispastic 28 29 30 31 32 33 34 35 36 37 Chronic LBP – Miscellaneous 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 35 of 41 https://mc.manuscriptcentral.com/bmj BMJ Page 68 of 74

1 2 3 Supplemental file 15. Forest plot adverse events 4 5 Acute LBP – Non-benzodiazepine antispasmodic 6 7 8 9 10 11 12 Confidential: For Review Only 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 Acute LBP – Antispastic 30 31 32 33 34 35 36 37 38 39 Acute LBP – Benzodiazepine 40 41 42 43 44 45 46 47 48 49 Chronic LBP – Miscellaneous 50 51 52 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 36 of 41 https://mc.manuscriptcentral.com/bmj Page 69 of 74 BMJ

1 2 3 Mixed LBP – Non-benzodiazepine antispasmodic 4 5 6 7 8 9 10 11 12 Confidential: For Review Only 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 37 of 41 https://mc.manuscriptcentral.com/bmj BMJ Page 70 of 74

1 2 3 Supplemental file 16. Forest plot serious adverse events 4 5 Acute LBP – Non-benzodiazepine antispasmodic 6 7 8 9 10 11 12 Confidential: For Review Only 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 38 of 41 https://mc.manuscriptcentral.com/bmj Page 71 of 74 BMJ

1 2 3 Supplemental file 17. Forest plot tolerability 4 5 Acute LBP – Non-benzodiazepine antispasmodic 6 7 8 9 10 11 12 Confidential: For Review Only 13 14 15 16 17 18 Acute LBP – Antispastic 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 39 of 41 https://mc.manuscriptcentral.com/bmj BMJ Page 72 of 74

1 2 3 Supplemental file 18. Forest plot dose subgroup analysis 4 5 Population: Acute low back pain 6 7 Medicine: Non-benzodiazepine antispasmodic 8 9 Outcome: Pain intensity 10 Follow-up: Immediate (≤ 2 weeks) 11 12 StandardConfidential: dose For Review Only 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 Above dose 36 37 38 39 40 41 42 43 44 45 46 Below dose 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cashin et al. 2021 Page 40 of 41 https://mc.manuscriptcentral.com/bmj Page 73 of 74 BMJ

1 2 3 Supplemental file 19. Sensitivity analyses for non-benzodiazepine antispasmodic medicines in acute LBP 4 5 Outcome Overall Removed trials Removed trials Removed trials Removed trials Removed Removed Removed trials 6 with an unclear measuring pain where measures for carisoprodol trials at trials with without a 7 definition for with a VRS of variance were high risk of data from trial placebo 8 Confidential:non-specific imputedFor Review bias Onlyregistry comparator 9 LBP record 10 11 (MD/RR [95% (MD/RR [95% (MD/RR [95% (MD/RR [95% (MD/RR [95% (MD/RR (MD/RR [95% (MD/RR [95% CI]; 12 CI]; Tau2; n] CI]; Tau2; n] CI]; Tau2; n] CI]; Tau2; n] CI]; Tau2; n] [95% CI]; CI]; Tau2; n] Tau2; n] 13 Tau2; n] 14 15 Pain intensity -7.7 (-12.1 to - -8.1 (-12.7 to - -9.7 (-15.4 to - -8.2 (-13.2 to - -8 (-14.3 to -1.7), 0.2 (-4.9 to -10.2 (-15.6 to -11 (-17 to -5.1), 16 (≤ 2 weeks) 3.3), 76.2, 3.6), 79.3, 3.9), 92.6, 3.2), 77.6, 103.9, n=1559 5.4), 0, -4.7), 96.4, 95.9, n=3488 17 n=4546 n=4450 n=2767 n=3495 n=672 n=2901 18 19 Acceptability 0.8 (0.6 to 1.1), 0.8 (0.6 to 1.1), 0.9 (0.6 to 1.3), 0.2 (0 to 0.8 (0.6 to 1), 0.1, 20 0.1, n=2834 0, n=2520 - - 0.2, n=1412 3.8), NA, - n=2332 21 n=30 22 23 Disability -3.3 (-7.3 to 0.7), -3.3 (-6.2 to -0.4), 2.3 (-3.6 to 8.3), 2.3 (-3.6 to -3.7 (-8.6 to -5.9 (-10.5 to - 24 (≤2 weeks) 20.2, n=2438 - - 4, n=2410 0, n=652 8.3), 0, 1.2), 26.7, 1.3), 17.5, n=1786 25 n=652 n=2020 26 27 Adverse 1.6 (1.2 to 2), 1.4 (1 to 2), 0.2, 1.2 (0.8 to 1.4 (1 to 2), 1.8 (1.3 to 2.4), events 0.1, n=3404 - - - n=1741 1.9), 0.1, 0.2, n=1741 0.1, n=2385 28 n=737 29 30 Tolerability 1.5 (0.6 to 3.5), 3.6 (0.9 to 14.7), 1.2 (0.5 to 3), 0.4, - - - - - 31 0.5, n=1641 0.1, n=254 n=1536 32 33 LBP, Low Back Pain; MD, Mean Difference; RR, Risk Ratio; CI, Confidence Interval; VRS, Verbal Rating Scale; NA, Not Applicable 34 35 36 37 38 39 40 41 42 Cashin et al. 2021 Page 41 of 41 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 74 of 74

1 2 3 4 5 6 7 8 9 10 11 Confidential: For Review Only 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 Figure 1. PRISMA flow diagram of the record selection process. 39 365x353mm (72 x 72 DPI) 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 75 of 74 BMJ

1 2 3 4 5 6 7 8 9 10 11 Confidential: For Review Only 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 Figure 2. Effect of non-benzodiazepine antispasmodic medicines compared to control on pain intensity (0- 46 100 scale) at immediate term (≤ 2 weeks) post randomisation for adults with LBP. Negative values for mean 47 differences indicate that effects favour muscle relaxant medicines compared to control. Whereas, negative values for trial observations indicate change from baseline. 48 49 324x451mm (72 x 72 DPI) 50 51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj