BMJ

Confidential: For Review Only Efficacy and safety of antidepressants for the treatment of spinal pain and osteoarthritis: A systematic review and meta-analysis

Journal: BMJ

Manuscript ID BMJ-2020-058600

Article Type: Research

BMJ Journal: BMJ

Date Submitted by the 18-May-2020 Author:

Complete List of Authors: Ferreira, Giovanni; The University of Sydney, Sydney School of Public Health, Faculty of Medicine and Health; Institute for Musculoskeletal Health McLachlan, Andrew; University of Sydney, Faculty of Pharmacy Lin, Chung-Wei; The University of Sydney, School of Public Health, Faculty of Medicine and Health; Institute for Musculoskeletal Health , University of Sydney Zadro, Joshua; University of Sydney School of Public Health; Institute for Musculoskeletal Health Abdel Shaheed, Christina; The University of Sydney, School of Public Health, Faculty of Medicine and Health; Institute for Musculoskeletal Health , University of Sydney O'Keeffe, Mary; The University of Sydney, School of Public Health, Faculty of Medicine and Health; Institute for Musculoskeletal Health , University of Sydney Maher, Christopher; The University of Sydney, School of Public Health, Faculty of Medicine and Health; Institute for Musculoskeletal Health

antidepressants, low back pain, osteoarthritis, Serotonin and Keywords: Noradrenaline Reuptake Inhibitors, Tricyclic, Pain, Pain management, Selective Serotonin Reuptake Inhibitor

https://mc.manuscriptcentral.com/bmj Page 1 of 56 BMJ

1 2 Efficacy and safety of antidepressants for the treatment of spinal 3 4 5 pain and osteoarthritis: A systematic review and meta-analysis 6 7 8 9 10 11 12 Confidential: For Review Only 13 14 Giovanni E Ferreira, phd candidate (0000-0002-8534-195X)1,2, Andrew J 15 16 McLachlan, professor (0000-0003-4674-0242)3, Chung-Wei Christine Lin, 17 18 professor (0000-0001-6192-7238)1,2, Joshua R Zadro, research fellow (0000- 19 20 0001-8981-2125),2, Christina Abdel-Shaheed (0000-0003-1258-5125), research 21 22 fellow 1,2, Mary O’Keeffe, research fellow (0000-0001-7104-9248) 1,2,4, Chris G 23 24 Maher, professor (0000-0002-1628-7857) 1,2 25 26 27 28 29 1. Sydney School of Public Health, Faculty of Medicine and Health, The University of Sydney, 30 31 Sydney, NSW, 2050, Australia 32 2. Institute for Musculoskeletal Health, Sydney, Australia 33 34 3. Sydney Pharmacy School, Faculty of Medicine and Health, University of Sydney, Sydney, NSW, 35 2050, Australia 36 37 4. School of Allied Health, Faculty of Education and Health Sciences, University of Limerick, 38 39 Limerick, V94 T9PX, Ireland 40 41 42 43 44 45 46 47 48 49 50 51 Abstract word count: 381 52 53 Manuscript word count: 2,990 54 55 Address all correspondence to: 56 Giovanni E Ferreira 57 58 School of Public Health, Faculty of Medicine and Health, The University of Sydney, Sydney, New 59 South Wales, Australia 60 PO Box M179, Missenden Road, Camperdown | NSW | 2050, Australia E-mail: [email protected] 1 https://mc.manuscriptcentral.com/bmj BMJ Page 2 of 56

1 2 ABSTRACT 3 4 5 Objectives: To investigate the efficacy and safety of antidepressants in patients with spinal pain and 6 7 osteoarthritis compared to placebo. 8 9 Design: Systematic review and meta-analysis 10 11 Data sources: MEDLINE, EMBASE, Cochrane Central Register of Controlled Trials, CINAHL, 12 Confidential: For Review Only 13 International Pharmaceutical Abstracts, ClinicalTrials.gov and the World Health Organization 14 15 International Clinical Trials Registry Platform from their inception to May 12, 2020. 16 17 Eligibility criteria: Randomised controlled trials investigating the efficacy and/or safety of any 18 antidepressant drug to placebo (active or inert) in participants with low back and neck pain, sciatica or 19 20 hip and/or knee osteoarthritis. Two independent reviewers selected studies. 21 22 Data Extraction and Synthesis: Two independent reviewers all the data. Pain and disability were 23 24 primary outcomes. Pain and disability scores were converted to a scale of 0 (no pain or disability) to 25 100 (worst pain or disability). We calculated weighted mean differences (MD) and 95% confidence 26 27 intervals (CI) using a random effects model. Safety (any adverse event, serious adverse events, and 28 29 proportion of participants dropping out of the trials due to adverse events were secondary outcomes). 30 We assessed risk of bias with the Cochrane Collaboration’s tool was used for assessing risk of bias. The 31 32 GRADE framework was used to assess the certainty of evidence. 33 34 Results: We included 33 randomised controlled trials (n = 5,110 participants). Moderate certainty 35 36 evidence showed that serotonin–norepinephrine reuptake inhibitors (SNRI) reduce pain in the short 37 term for participants with spinal pain (MD -5.3, 95% CI -7.3 to -3.3) and osteoarthritis (MD -4.7, 95% 38 39 CI -6.3 to -3.1). Very low certainty evidence showed that SNRI antidepressants reduced sciatica pain 40 in the immediate (MD -18.6, 95% CI -31.9 to -5.3), but not in the short-term; and low to very low 41 42 certainty evidence showed that tricyclic antidepressants (TCA) did not reduce sciatica pain in the 43 immediate (MD -0.9, 95% CI -5.4 to 3.6), but did it so in the short and intermediate term. SNRI reduced 44 45 disability for spinal pain in the short-term (MD -3.6, 95% CI -5.2 to -1.9) and osteoarthritis in the 46 47 immediate and short-term. 48 49 Conclusions: There is moderate certainty evidence that SNRI antidepressants reduce pain and disability 50 in spinal pain and osteoarthritis, but the effects are small and unlikely to be clinically relevant. TCA 51 52 and SNRI antidepressants may be effective for sciatica, but the certainty of evidence ranged from low 53 54 to very low. 55 56 57 58 59 60

2 https://mc.manuscriptcentral.com/bmj Page 3 of 56 BMJ

1 2 INTRODUCTION 3 4 5 Spinal pain (low back pain or neck pain with or without radicular symptoms) and osteoarthritis are 6 1 7 leading causes of disability worldwide. At any given time, low back pain and neck pain affect 7.3% 8 and 5% of the population, respectively, whereas symptomatic osteoarthritis affects 12% of the 9 10 population.2,3 In 2016, spinal pain and osteoarthritis accounted for $214.5 billion in health care spending 11 in the United States in 2016, with spinal pain having the highest expenditure among all health 12 Confidential: For Review Only 13 conditions.4 14 15 Prescription of antidepressants is increasing worldwide for a range of indications.5 Among medicines 16 17 associated with dependence and withdrawal, antidepressants are the most commonly prescribed 18 medicine in the United Kingdom, with more people being prescribed antidepressants (7.2 million) than 19 20 opioid analgesics (5.6 million).6 The use of antidepressants to treat pain, especially chronic pain, is also 21 22 very common. For example, antidepressants are the fourth most commonly prescribed medicine for low 23 back pain in the United States, with over a quarter of people with chronic low back pain being prescribed 24 25 an antidepressant within 3 months of first diagnosis.7 In Canada, tricyclic antidepressants amitriptyline 26 and nortriptyline are primarily prescribed for pain, representing 48.4% and 57.4% of all prescriptions 27 28 for these antidepressants, respectively.8 In the United Kingdom, 16% of all antidepressant prescriptions 29 were for pain in children and adolescents, and their rate of prescription has almost tripled from 2003 to 30 31 2013.9 The widespread use of antidepressants for pain is also seen in middle-income countries. 32 33 Amitriptyline is widely used in primary care in South Africa, and osteoarthritis is the most common 34 diagnosis leading to its prescription, representing just over a quarter of all amitriptyline prescribed.10 35 36 Antidepressants are endorsed by most evidence-based clinical practice guidelines for low back pain 37 38 (75% of guidelines),11 and two recently published osteoarthritis guidelines.12,13 For example, the 39 40 American College of Physicians recommend serotonin and norepinephrine reuptake inhibitor (SNRI) 41 duloxetine for low back pain.14 The National Institute for Health and Care Excellence (NICE) 42 43 recommend amitriptyline or duloxetine as first-line therapy for people with different forms of 44 neuropathic pain, including spinal pain with radicular symptoms.15 Osteoarthritis guidelines, such as 45 46 those from Osteoarthritis Research Society International (OARSI), and the American College of 47 48 Rheumatology recommend the use of duloxetine in pain management of this condition. In the OARSI 49 guideline, the recommendation targets people with osteoarthritis with concomitant depression and/or 50 12,13 51 widespread pain. Evidence supporting the use of antidepressants is uncertain; existing systematic 52 reviews of antidepressants for spinal pain and osteoarthritis are either outdated16,17 or consider only one 53 54 type of antidepressant (eg duloxetine) and did not assessed the certainty of evidence.18 None of the 55 existing reviews included unpublished records from trial registries, a practice that may overestimate the 56 57 efficacy of pharmacological interventions for pain control.19 The aim of this systematic review is to 58 59 address these knowledge gaps by investigating the efficacy and safety of antidepressants in patients 60 with spinal pain (including sciatica) or hip or knee osteoarthritis.

3 https://mc.manuscriptcentral.com/bmj BMJ Page 4 of 56

1 2 METHODS 3 4 Data sources and searches 5 6 This review was prospectively registered20 and reported according to the PRISMA reporting 7 8 guidelines.22 We searched MEDLINE, EMBASE, Cochrane Central Register of Controlled Trials, 9 10 CINAHL, International Pharmaceutical Abstracts, ClinicalTrials.gov and the World Health 11 Organization International Clinical Trials Registry Platform from their inception to May 12, 2020 12 Confidential: For Review Only 13 (Supplemental file 1). Two authors (GEF, MOK) independently screened records by titles and abstracts. 14 Two authors (GEF, JZ) read full-texts of potentially eligible studies to determine eligibility (see 15 16 supplemental file 2 for a list of excluded trials with reasons). Any disagreements were resolved by 17 consensus. 18 19 20 Eligibility criteria 21 22 We included randomised controlled trials comparing any antidepressant drug to placebo in participants 23 24 with spinal pain (neck or low back pain with or without radicular symptoms) or hip and/or knee 25 osteoarthritis. Symptoms of any duration were included. Trials including drug combinations were 26 27 eligible if the treatment contrast between groups was antidepressant versus placebo. The placebo 28 comparator could be active (a substance that has no known effect on pain but may mimic the adverse 29 30 effects of antidepressants) or inert (a substance that is not thought to have a therapeutic or adverse 31 effect). We included reports published in peer-reviewed journals as well as unpublished data posted on 32 33 trial registry platforms (e.g. ClinicalTrials.gov). Trials reporting data on either pain, disability or 34 35 adverse events were included. We placed no restrictions on language or publication date. We excluded 36 studies including participants with serious spinal pathologies (e.g. axial spondyloarthritis, fractures, 37 38 cancer) and rheumatic conditions (e.g. rheumatoid arthritis), unless data for spinal pain or knee and/or 39 hip osteoarthritis were reported separately. Studies where participants received previous spinal or 40 41 osteoarthritis surgery were eligible, but studies evaluating immediate post-operative pain management 42 43 (i.e. surgery within past month) were excluded. Conference abstracts were excluded. 44 45 Data extraction 46 47 Two authors (GEF, JZ) independently extracted data. Whenever possible, we extracted post-treatment 48 49 means, standard deviations and number of participants per group for each outcome. In the absence of 50 post scores, we extracted data according to the following hierarchy: between group differences at 51 52 follow-up and pre to post within group change scores. When variability measures were not reported, we 53 used estimation methods recommended by the Cochrane Handbook for Systematic Reviews of 54 55 Interventions version 6.0 (see Supplemental files 3 to 7 for details on how effect sizes were calculated

56 21 57 for each outcome measure within each study). 58 59 Outcomes 60

4 https://mc.manuscriptcentral.com/bmj Page 5 of 56 BMJ

1 2 The primary outcomes were pain intensity and disability. Adverse events were a secondary outcome. 3 4 Adverse events included the number of participants (i) reporting any adverse event (as defined by each 5 study), (ii) reporting any serious adverse event (as defined by each study); and (iii) who withdrew due 6 7 to adverse effects. 8 9 Risk of bias and certainty of evidence 10 11 Two reviewers (GEF, MOK) rated risk of bias of trials using the Cochrane Collaboration Risk of Bias 12 Confidential: For Review Only 13 Tool.22 Disagreements were resolved by consensus. We assessed the certainty of evidence using the 14 Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework.23 We 15 16 downgraded the certainty of evidence by one level if a serious flaw was present in any of the following 17 domains: limitations in study design, inconsistency, imprecision, small study bias. We did not 18 19 downgrade for indirectness since patients, interventions and comparators were similar across 20 21 comparisons (See supplemental file 8 for a description of the GRADE framework used). 22 23 Data synthesis and analysis 24 25 Outcomes were classified into immediate term (≤2 weeks); short term (>2 weeks but ≤ 3 months); 26 27 intermediate term (>3 months but ≤12 months); long term (>12 months). In studies with multiple time 28 points, we extracted data from the time point closest to 2 weeks, 3, 6 and 12 months. We converted pain 29 30 and disability scores to a common 0-100 scale, where 0 means no pain or disability and 100 means the 31 worst possible pain or disability. For studies that reported multiple pain measures (e.g. “average pain”, 32 33 “worst pain”, and “least pain”), we extracted the “average pain” score where possible.24 Weighted mean 34 35 differences (MD) (95% confidence interval (CI)) were calculated for continuous outcome measures, 36 and risk ratios (95% CI) were calculated for dichotomous outcomes. Statistical heterogeneity in each 37 38 meta-analysis was determined by means of the I2 test. We used a random-effects model across all 39 comparisons. We grouped antidepressants based on their pharmacological class. Between-group mean 40 41 differences below 10 points for pain and disability were small and not clinically important. This

42 17,25 43 threshold has been used in other reviews of pharmacological treatments for spinal pain, and is also 44 the recommended threshold for pain and disability in osteoarthritis.26,27 We used Comprehensive Meta- 45 46 Analysis version 3.3 (Englewood, NJ, USA). 47 48 Exploratory meta-regression 49 50 We performed meta-regression to explore the moderator effects of risk of bias (high risk of bias if at 51 52 least one domain classified as high risk of bias or more than half of the bias domains classified as 53 unclear), industry funding (yes/no), small study effects (< 100 participants per arm), depression listed 54 55 as inclusion criteria (yes/no), and dose of antidepressant on pain. We grouped all antidepressant classes 56 57 together for this analysis. 58 59 60

5 https://mc.manuscriptcentral.com/bmj BMJ Page 6 of 56

1 2 RESULTS 3 4 Our search retrieved 2,216 records, 1,930 of which were screened after duplicates were removed. We 5 6 excluded 1,795 records based on titles and abstracts. Out of 135 potentially relevant trials screened for 7 eligibility, we included 33 trials enrolling 5,110 participants. 8 9 10 Most trials (n = 28, 84.9%) used a parallel-group design, whereas five trials (15.2%) used a cross-over 11 design with washout periods ranging from one to two weeks.28-32 Fourteen trials were industry- 12 Confidential: For Review Only 13 sponsored24,31,33-44 and funding source was unclear in five trials. All but one trial reported data from 14 participants with chronic pain.45 Only six trials restricted inclusion to participants with depression.37- 15 16 39,42,45,59 Two other trials had between 40% and 50% of the sample with depression.39,40 Others either 17 excluded participants with depressive disorder or did not mention depression in the eligibility criteria. 18 19 The median duration of the drug regimen was 8 weeks. 20 21 Serotonin-norepinephrine reuptake inhibitor (SNRI) antidepressants was the most studied 22 23 antidepressant class (15 trials), followed by tricyclic antidepressants (TCA, 14 trials), serotonin 24 reuptake inhibitor (SSRI, 3 trials), norepinephrine-dopamine reuptake inhibitor (NDRI, 1 trial), 25 26 Serotonin antagonist and reuptake inhibitors (SARI, 1 trial), and tetracyclic antidepressants (1 trial). 27 28 (Supplemental file 9). 29 30 Twenty-five trials had at least one domain classified as high risk of bias. Twenty-six trials were unclear 31 in describing the methods used to conceal allocation, and therefore were at unclear risk of selection 32 33 bias. One trial was at high risk of performance bias due to inadequate blinding of participants, and 13 34 35 were at unclear risk of performance bias. One trial was at high risk of detection bias due to inadequate 36 blinding of participants, and another 12 trials were at unclear risk of detection bias. Eighteen trials were 37 38 at high risk of attrition bias. Five and 14 trials were at high and unclear risk of reporting bias, 39 respectively (Supplemental file 10). 40 41 42 Spinal pain 43 44 Nineteen trials (23 comparisons) enrolling 2,698 participants determined the efficacy of antidepressants 45 for spinal pain. Of these, 16 trials (n = 2,388 participants) reported data for low back pain, 1 trial 46 47 reported data for neck pain (n = 220 participants), 2 trials reported data for low back and neck pain (n 48 49 = 90). Only one trial of TCA evaluated intermediate-term outcomes. No trials evaluated long-term 50 outcomes, and only pain outcomes were measured at the immediate term follow-up. 51 52 There is moderate certainty evidence that SNRI antidepressants reduced pain in the immediate (MD - 53 54 3.7, 95% CI -5.9 to -1.4; 7 trials, 1,305 participants) and short-term (MD -5.3, 95% CI -7.3 to -3.3; 4 55 56 trials, 1,415 participants) (Figure 2). SNRI antidepressants also reduced disability in the short term (MD 57 -3.5, 95% CI -5.2 to -1.9; 4 trials, 1,423 participants) (Figure 3) The effect of SNRI was small and 58 59 below our pre-determined threshold of clinical importance. Evidence ranging from low to very low 60

6 https://mc.manuscriptcentral.com/bmj Page 7 of 56 BMJ

1 2 certainty showed no benefit of a range of antidepressant classes, including SSRI, TCA, Tetracyclic, 3 4 SARI and NDRI antidepressants for pain and disability. 5 6 We conducted a post-hoc sensitivity analysis to explore the effect of removing one trial from the pooled 7 estimates for spinal pain.46 Participants in this trial had neck pain, received a dose of amitriptyline (5 8 9 mg/day) which is much lower than the minimum dose recommended for pain relief (10 mg/day to 25 10 mg/day), and reported unexpectedly large improvements in pain and disability compared to other 11 12 studies. TheConfidential: exclusion of this trial fully explained For the Review heterogeneity for disabilityOnly and 90% for pain. 13 14 After excluding this trial, TCA antidepressants were found to significantly reduce pain (MD -5.4 95% 15 CI -10.0 to -0.8) and disability (MD -6.5, 95% CI -10.2 to -2.9), but effects were still small and below 16 17 our pre-determined threshold for clinical importance. 18 19 Sciatica 20 21 Six trials enrolling 267 participants reported data for sciatica. No trials evaluated long-term outcomes. 22 23 24 There is very low certainty evidence that SNRI antidepressants reduced pain in the immediate (MD - 25 18.6, 95% CI -31.9 to -5.3; 1 trial, 50 participants), but not in the short-term (MD -17.5, 95% CI -42.9 26 27 to 7.9; 3 trials, 96 participants). Low to very low certainty evidence showed that TCA did not reduce 28 pain in the immediate-term (MD -0.9, 95% CI -5.4 to 3.6; 3 trials, 145 participants), but did it so in the 29 30 short (MD -16.0, 95% CI -31.5 to -0.4; 2 trials, 116 participants) and intermediate (MD -27.0, 95% CI 31 -36.1 to -17.9; 1 trial, 60 participants) terms (Figure 2). TCA did not reduce disability in the immediate 32 33 and short term but did so in the intermediate term (Table 3). 34 35 Osteoarthritis 36 37 38 Eight trials (8 comparisons) enrolling 2,145 participants evaluated the efficacy of antidepressants in 39 participants with knee osteoarthritis. All trials were of SNRI antidepressants. None of the osteoarthritis 40 41 trials included participants with hip osteoarthritis or evaluated intermediate or long-term outcomes. 42 43 There is moderate certainty evidence that SNRI antidepressants reduced pain in the immediate (MD - 44 45 4.7, 95% CI -6.3 to -3.1; 4 trials, 1,328 participants) and short term (MD -4.7, 95% CI -11.7 to -6.8; 8 46 trials, 1,941 participants) (Figure 2). Low certainty evidence showed that SNRI antidepressants reduced 47 48 disability in the immediate (MD -5.1, 95% CI -7.3 to -2.9; 1 trial, 353 participants) and short term (MD 49 -6.0, 95% CI -8.1 to -3.9; 7 trials, 1,810 participants) (Figure 3). The effect of SNRI was small and 50 51 below our pre-determined threshold of clinically importance. 52 53 Safety 54 55 Twenty-one trials (25 comparisons) enrolling 4,107 participants determined the safety of 56 57 antidepressants for spinal pain and osteoarthritis. The nature and reporting of adverse events varied 58 59 markedly across trials. Nausea was the most reported adverse event, followed by dry mouth and 60 sleepiness.

7 https://mc.manuscriptcentral.com/bmj BMJ Page 8 of 56

1 2 There is low certainty evidence that SNRI antidepressants increase the risk of any adverse event (62.8% 3 4 vs 49.8%; RR 1.2, 95% CI 1.1 to 1.3; 13 trials, 3,436 participants), but not serious adverse events (1.7% 5 vs 1.3%, RR 1.1, 95% CI 0.6 to 2.1; 10 trials, 3,297 participants) (Figure 4). Participants taking SNRI 6 7 antidepressants were also more likely to drop out of the study due to adverse events (12.3% vs 5.2%; 8 RR 2.2, 95% CI 1.7 to 2.7). None of the other antidepressant classes increased the risk of adverse events 9 10 or drop-outs due to adverse events. 11 12 ExploratoryConfidential: sensitivity analyses For Review Only 13 14 There was no interaction between risk of bias, study size, industry funding, depression diagnosis and 15 16 daily dosage of duloxetine and treatment effects (Supplemental file 11). 17 18 DISCUSSION 19 20 Statement of principal findings 21 22 23 This review found moderate certainty evidence that SNRI antidepressants reduce pain and disability in 24 people with spinal pain or osteoarthritis. Low certainty evidence showed that TCA was ineffective for 25 26 pain and disability for spinal pain. TCA and SNRI antidepressants may reduce pain in participants with 27 sciatica. Only SNRI antidepressants significantly increased the risk of adverse events, but there is low 28 29 to very low certainty across antidepressant classes. 30 31 Strengths and weaknesses 32 33 Our review has several strengths. We registered this review prospectively and performed a 34 35 comprehensive literature search, including searches on clinical trial registries. We also complied with 36 37 all items proposed by the AMSTAR-2 checklist, meaning that our review can be considered of high 38 methodological quality.47 Our review updates the evidence for spinal pain, sciatica and osteoarthritis. 39 40 For example for spinal pain and sciatica we included 25 trials (n = 2,965 participants) whereas the 41 previous review included 16 trials.17 While previous review had not been able to estimate an effect of 42 43 antidepressants for sciatica, we were able to pool data for TCA and SNRI antidepressants for that

44 16,17 45 condition (six trials). Our review also updates the evidence for osteoarthritis. We pooled data from 46 eight osteoarthritis trials whereas a previous review only pooled five trials.18 47 48 Our review has limitations. There is marked uncertainty in the effects for sciatica and safety outcomes. 49 50 Second, we were not able to explore a dose-response relationship for most antidepressants due to the 51 52 low number of studies and varied doses. Nevertheless, we were able to conduct such analysis for 53 duloxetine (14 trials) and found that 60 mg/day produced similar effects on pain to 120 mg/day. 54 55 Meaning of the study 56 57 58 While SNRI antidepressants reduce pain and disability for spinal pain or osteoarthritis the effects may 59 be too small to be considered worthwhile by patients. Our threshold for clinical importance (>10 points 60 on a 0-100 scale) was pre-specified and generous as others have proposed higher thresholds of >15 and 8 https://mc.manuscriptcentral.com/bmj Page 9 of 56 BMJ

1 2 >20 points).48 However, for sciatica the effects of both TCA and SNRI antidepressants exceeded this 3 4 threshold but were based on less certain evidence. The findings on sciatica are in accordance with a 5 systematic review that reported moderate to strong certainty evidence for clinically important effects of 6 49 7 TCAs and SNRI antidepressants on neuropathic pain. However, caution is needed in interpreting our 8 findings for sciatica. All sciatica trials were small, had imprecise estimates and were at high risk of bias, 9 10 which reduced the certainty of evidence to low and very low. Hence further trials are needed before 11 firm conclusions about the effects of antidepressants for sciatica can be made. 12 Confidential: For Review Only 13 14 Participants diagnosed with depression did not benefit more from antidepressants for pain compared to 15 those without depression. In both subgroups, improvements were below the threshold for clinically 16 17 important effects. Further, none of the trials that contributed data for the subgroup with depression were 18 for osteoarthritis. The 2019 OARSI guideline recommends antidepressants (duloxetine) for knee 19 20 osteoarthritis comorbid with depression and/or widespread pain disorders.12 Some of the trials included 21 22 in our review explored the indirect effect of the antidepressants on depression and its role in mediating 23 the effect on pain with path analysis.33,42,43 However, these trials excluded participants with major 24 25 depressive disorder, and participants had very low average scores (e.g. around 5 points43) on the Beck 26 Depression Inventory-II, which were within the minimal range and below the proposed thresholds for 27 28 diagnosing depression.50 Because usually doses used to treat pain are below those used for the treatment 29 of depression, more research into the moderating role of depression on the effects of antidepressants on 30 31 pain is needed – particularly in osteoarthritis trials. 32 33 CONCLUSION 34 35 36 Moderate certainty evidence showed that the effect of SNRI antidepressants on pain and disability in 37 participants with spinal pain and osteoarthritis is small and may not be considered worthwhile by 38 39 patients. TCA and SNRI antidepressants may be effective for sciatica, but the certainty of evidence 40 ranged from low to very low. SNRI antidepressants slightly increasing the risk of adverse events, but 41 42 not serious adverse events, although the certainty of the evidence was low. 43 44 Contributors 45 46 GF, AM, CL, JZ, MK, CS, and CM designed the review protocol. GF, JZ, MK developed the search 47 48 strategy and performed study selection. GF, JZ, MK, CS extracted data. GF analysed the data. GF 49 50 drafted the manuscript and JZ, MK, and CS contributed to the drafting of the review. AM, CL and CM 51 revised it critically for important intellectual content. All authors approved the final version of the 52 53 article. All authors had access to all the data in the study and can take responsibility for the integrity of 54 the data and the accuracy of the data analysis. GF is guarantor. 55 56 57 Funding 58 This research received no specific grant from any funding agency in the public, commercial, or not-for- 59 60 profit sectors. GF hold a PhD scholarship from Coordenação de Aperfeiçoamento de Pessoal de Nível

9 https://mc.manuscriptcentral.com/bmj BMJ Page 10 of 56

1 2 Superior (CAPES), Brazil. Prof Christine Lin is supported by an NHMRC fellowship (APP1061400). 3 4 Chris Maher is supported by a Principal Research Fellowship from Australia’s National Health and 5 Medical Research Council (NHMRC) (APP1103022) as well as a Program grant (APP1113532) and 6 7 Centre for Research Excellence grant (APP1134856). 8 9 Competing interests declaration 10 11 The Sydney Pharmacy School receives funding from GlaxoSmithKline for a postgraduate scholarship 12 Confidential: For Review Only 13 supervised by AM. CL is supported by an NHMRC fellowship (APP1061400), conducted an 14 investigator initiated NHMRC funded trial (PRECISE) published in NEJM in 2017, for which Pfizer 15 16 Australia provided investigational product only, and currently conducts an investigator initiated 17 NHMRC of opioids for acute low back pain (OPAL). All other authors have nothing to disclose. CM 18 19 has received research grants from various government and not for profit agencies. Flexeze provided 20 21 heat wraps at no cost for the SHaPED trial that he is an investigator on. His expenses have been covered 22 by professional associations hosting conferences he has spoken at. 23 24 Ethical approval 25 26 27 Not required 28 29 Data sharing 30 31 No additional data is available 32 33 Transparency 34 35 36 The lead author (GF) affirms that the manuscript is an honest, accurate, and transparent account of the 37 study being reported; no important aspects of the study have been omitted. 38 39 40 Patient and Public involvement 41 Patients or the public WERE NOT involved in the design, or conduct, or reporting, or dissemination 42 plans of our research. 43 44 Dissemination declaration 45 46 We have plans to disseminate our findings to patient organisations 47 48 49 50 51 52 53 54 55 56 57 58 59 60

10 https://mc.manuscriptcentral.com/bmj Page 11 of 56 BMJ

1 2 REFERENCES 3 4 1. Disease GBD, Injury I, Prevalence C. Global, regional, and national incidence, prevalence, and 5 years lived with disability for 354 diseases and injuries for 195 countries and territories, 6 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet. 7 2018;392(10159):1789-1858. 8 2. Cross M, Smith E, Hoy D, et al. The global burden of hip and knee osteoarthritis: estimates 9 from the global burden of disease 2010 study. Ann Rheum Dis. 2014;73(7):1323-1330. 10 3. Hartvigsen J, Hancock MJ, Kongsted A, et al. What low back pain is and why we need to pay 11 12 attention.Confidential: Lancet. 2018;391(10137):2356-2367. For Review Only 13 4. Dieleman JL, Cao J, Chapin A, et al. US Health Care Spending by Payer and Health Condition, 14 1996-2016. JAMA. 2020;323(9):863-884. 15 5. OECD. Antidepressant drugs consumption, 2000 and 2015 (or nearest year). 2017. 16 6. Marsden J, White M, Annand F, et al. Medicines associated with dependence or withdrawal: 17 a mixed-methods public health review and national database study in England. Lancet 18 Psychiatry. 2019;6(11):935-950. 19 7. Ivanova JI, Birnbaum HG, Schiller M, Kantor E, Johnstone BM, Swindle RW. Real-world 20 practice patterns, health-care utilization, and costs in patients with low back pain: the long 21 22 road to guideline-concordant care. Spine J. 2011;11(7):622-632. 23 8. Wong J, Motulsky A, Abrahamowicz M, Eguale T, Buckeridge DL, Tamblyn R. Off-label 24 indications for antidepressants in primary care: descriptive study of prescriptions from an 25 indication based electronic prescribing system. BMJ. 2017;356:j603. 26 9. John A, Marchant AL, Fone DL, et al. Recent trends in primary-care antidepressant 27 prescribing to children and young people: an e-cohort study. Psychol Med. 28 2016;46(16):3315-3327. 29 10. Coetzee R, Johnson Y, van Niekerk J, Namane M. Amitriptyline prescribing in public sector 30 31 healthcare facilities in the Western Cape, South Africa. PLoS ONE. 2020;15(4):e0231675. 32 11. Oliveira CB, Maher CG, Pinto RZ, et al. Clinical practice guidelines for the management of 33 non-specific low back pain in primary care: an updated overview. Eur Spine J. 34 2018;27(11):2791-2803. 35 12. Bannuru RR, Osani MC, Vaysbrot EE, et al. OARSI guidelines for the non-surgical 36 management of knee, hip, and polyarticular osteoarthritis. Osteoarthritis Cartilage. 37 2019;27(11):1578-1589. 38 13. Kolasinski SL, Neogi T, Hochberg MC, et al. 2019 American College of Rheumatology/Arthritis 39 40 Foundation Guideline for the Management of Osteoarthritis of the Hand, Hip, and Knee. 41 Arthritis Care Res (Hoboken). 2020;72(2):149-162. 42 14. Qaseem A, Wilt TJ, McLean RM, Forciea MA, Clinical Guidelines Committee of the American 43 College of P. Noninvasive Treatments for Acute, Subacute, and Chronic Low Back Pain: A 44 Clinical Practice Guideline From the American College of Physicians. Ann Intern Med. 45 2017;166(7):514-530. 46 15. National Institute for Health and Clinical Excellence (NICE). Neuropathic pain in adults: 47 pharmacological management in non-specialist settings. In: NICE, ed2013. 48 16. Pinto RZ, Maher CG, Ferreira ML, et al. Drugs for relief of pain in patients with sciatica: 49 50 systematic review and meta-analysis. BMJ. 2012;344:e497. 51 17. Chou R, Deyo R, Friedly J, et al. Systemic Pharmacologic Therapies for Low Back Pain: A 52 Systematic Review for an American College of Physicians Clinical Practice Guideline. Ann 53 Intern Med. 2017;166(7):480-492. 54 18. Osani MC, Bannuru RR. Efficacy and safety of duloxetine in osteoarthritis: a systematic 55 review and meta-analysis. Korean J Intern Med. 2019;34(5):966-973. 56 19. Bagg MK, O’Hagan E, Zahara P, et al. Reviews may overestimate the effectiveness of 57 medicines for back pain: systematic review and meta-analysis. J Clin Epidemiol. 58 59 20. Ferreira G, Lin C-WC, Zadro JR, et al. Efficacy and safety of antidepressants for the treatment 60 of low back pain and osteoarthritis: protocol for a systematic review and meta-analysis. MedRxiv. 2020:2020.2004.2030.20086827. 11 https://mc.manuscriptcentral.com/bmj BMJ Page 12 of 56

1 2 21. Higgins JPT TJ, Chandler J, Cumpston M, Li T, Page MJ, Welch VA (editors). Cochrane 3 Handbook for Systematic Reviews of Interventions version 6.0 (updated July 2019). Cochrane; 4 2019. 5 22. Higgins JPT, Altman DG, Gøtzsche PC, et al. The Cochrane Collaboration’s tool for assessing 6 risk of bias in randomised trials. BMJ. 2011;343:d5928. 7 8 23. Schünemann H BJ, Guyatt G, Oxman A. GRADE Handbook: Introduction to GRADE Handbook. 9 2013. 10 24. Wang G, Bi L, Li X, et al. Efficacy and safety of duloxetine in Chinese patients with chronic 11 pain due to osteoarthritis: a randomized, double- blind, placebo-controlled study. 12 OsteoarthritisConfidential: Cartilage. 2017;25(6):832-838. For Review Only 13 25. Machado GC, Maher CG, Ferreira PH, Day RO, Pinheiro MB, Ferreira ML. Non-steroidal anti- 14 inflammatory drugs for spinal pain: a systematic review and meta-analysis. Ann Rheum Dis. 15 2017;76(7):1269. 16 17 26. Reginster JY, Reiter-Niesert S, Bruyere O, et al. Recommendations for an update of the 2010 18 European regulatory guideline on clinical investigation of medicinal products used in the 19 treatment of osteoarthritis and reflections about related clinically relevant outcomes: expert 20 consensus statement. Osteoarthritis Cartilage. 2015;23(12):2086-2093. 21 27. Bellamy N, Buchanan WW, Goldsmith CH, Campbell J, Stitt LW. Validation study of WOMAC: 22 a health status instrument for measuring clinically important patient relevant outcomes to 23 antirheumatic drug therapy in patients with osteoarthritis of the hip or knee. J Rheumatol. 24 1988;15(12):1833-1840. 25 28. Schliessbach J, Siegenthaler A, Butikofer L, et al. Effect of single-dose imipramine on chronic 26 27 low-back and experimental pain. A randomized controlled trial. PLoS One. 28 2018;13(5):e0195776. 29 29. Schukro RP, Oehmke MJ, Geroldinger A, Heinze G, Kress HG, Pramhas S. Efficacy of 30 Duloxetine in Chronic Low Back Pain with a Neuropathic Component: A Randomized, 31 Double-blind, Placebo-controlled Crossover Trial. Anesthesiology. 2016;124(1):150-158. 32 30. Khoromi S, Cui L, Nackers L, Max MB. Morphine, nortriptyline and their combination vs. 33 placebo in patients with chronic lumbar root pain. Pain. 2007;130(1-2):66-75. 34 31. Katz J, Pennella-Vaughan J, Hetzel RD, Kanazi GE, Dworkin RH. A randomized, placebo- 35 36 controlled trial of bupropion sustained release in chronic low back pain. J Pain. 37 2005;6(10):656-661. 38 32. Pheasant H, Bursk A, Goldfarb J, Azen SP, Weiss JN, Borelli L. Amitriptyline and Chronic Low- 39 Back-Pain - a Randomized Double-Blind Crossover Study. Spine (Phila Pa 1976). 40 1983;8(5):552-557. 41 33. Konno S, Oda N, Ochiai T, Alev L. Randomized, Double-blind, Placebo-controlled Phase III 42 Trial of Duloxetine Monotherapy in Japanese Patients With Chronic Low Back Pain. Spine 43 (Phila Pa 1976). 2016;41(22):1709-1717. 44 45 34. Marks DM, Pae CU, Patkar AA. A double-blind, placebo-controlled, parallel-group pilot study 46 of milnacipran for chronic radicular pain (sciatica) associated with lumbosacral disc disease. 47 Prim Care Companion CNS Disord. 2014;16(4). 48 35. Skljarevski V, Zhang S, Desaiah D, et al. Duloxetine versus placebo in patients with chronic 49 low back pain: a 12-week, fixed-dose, randomized, double-blind trial. J Pain. 50 2010;11(12):1282-1290. 51 36. Skljarevski V, Ossanna M, Liu-Seifert H, et al. A double-blind, randomized trial of duloxetine 52 versus placebo in the management of chronic low back pain. Eur J Neurol. 2009;16(9):1041- 53 1048. 54 55 37. Skljarevski V, Desaiah D, Liu-Seifert H, et al. Efficacy and safety of duloxetine in patients with 56 chronic low back pain. Spine (Phila Pa 1976). 2010;35(13):E578-585. 57 38. Dickens C, Jayson M, Sutton C, Creed F. The relationship between pain and depression in a 58 trial using paroxetine in sufferers of chronic low back pain. Psychosomatics. 2000;41(6):490- 59 499. 60

12 https://mc.manuscriptcentral.com/bmj Page 13 of 56 BMJ

1 2 39. Goodkin K, Gullion CM, Agras WS. A Randomized, Double-Blind, Placebo-Controlled Trial of 3 Trazodone Hydrochloride in Chronic Low-Back-Pain Syndrome. J Clin Psychopharm. 4 1990;10(4):269-278. 5 40. Jenkins DG, Ebbutt AF, Evans CD. Tofranil in the treatment of low back pain. J Int Med Res. 6 1976;4(2 Suppl):28-40. 7 8 41. Uchio Y, Enomoto H, Alev L, et al. A randomized, double-blind, placebo-controlled Phase III 9 trial of duloxetine in Japanese patients with knee pain due to osteoarthritis. J Pain Res. 10 2018;11:809-821. 11 42. Chappell AS, Ossanna MJ, Liu-Seifert H, et al. Duloxetine, a centrally acting analgesic, in the 12 treatmentConfidential: of patients with osteoarthritis For knee Review pain: A 13-week, randomized, Only placebo- 13 controlled trial. Pain. 2009;146(3):253-260. 14 43. Chappell AS, Desaiah D, Liu-Seifert H, et al. A Double-blind, Randomized, Placebo-controlled 15 Study of the Efficacy and Safety of Duloxetine for the Treatment of Chronic Pain Due to 16 17 Osteoarthritis of the Knee. Pain Pract. 2011;11(1):33-41. 18 44. Frakes EP, Risser RC, Ball TD, Hochberg MC, Wohlreich MM. Duloxetine added to oral 19 nonsteroidal anti-inflammatory drugs for treatment of knee pain due to osteoarthritis: 20 results of a randomized, double-blind, placebo-controlled trial. Curr Med Res Opin. 21 2011;27(12):2361-2372. 22 45. Pirbudak L, Karakurum G, Satana T, et al. Epidural steroid injection and amitriptyline in the 23 management of acute low back pain originating from lumbar disc herniation. Joint Dis Rel 24 Surg. 2003;14(2):89-93. 25 46. Maarrawi J, Hay JA, Kobaiter-Maarrawi S, Tabet P, Peyron R, Garcia-Larrea L. Randomized 26 27 double-blind controlled study of bedtime low-dose amitriptyline in chronic neck pain. Eur J 28 Pain. 2018;22(6):1180-1187. 29 47. Shea BJ, Reeves BC, Wells G, et al. AMSTAR 2: a critical appraisal tool for systematic reviews 30 that include randomised or non-randomised studies of healthcare interventions, or both. 31 BMJ. 2017;358:j4008. 32 48. Olsen MF, Bjerre E, Hansen MD, Tendal B, Hilden J, Hróbjartsson A. Minimum clinically 33 important differences in chronic pain vary considerably by baseline pain and methodological 34 factors: systematic review of empirical studies. J Clin Epidemiol. 2018;101:87-106.e102. 35 36 49. Finnerup NB, Attal N, Haroutounian S, et al. Pharmacotherapy for neuropathic pain in adults: 37 a systematic review and meta-analysis. The Lancet Neurology. 2015;14(2):162-173. 38 50. Smarr KL, Keefer AL. Measures of depression and depressive symptoms: Beck Depression 39 Inventory-II (BDI-II), Center for Epidemiologic Studies Depression Scale (CES-D), Geriatric 40 Depression Scale (GDS), Hospital Anxiety and Depression Scale (HADS), and Patient Health 41 Questionnaire-9 (PHQ-9). Arthritis Care Res (Hoboken). 2011;63 Suppl 11:S454-466. 42 51. Gould HM, Atkinson JH, Chircop-Rollick T, et al. A randomized placebo-controlled trial of 43 desipramine, cognitive behavioral therapy, and active placebo therapy for low back pain. 44 45 Pain. 2020. 46 52. Urquhart DM, Wluka AE, van Tulder M, et al. Efficacy of Low-Dose Amitriptyline for Chronic 47 Low Back Pain: A Randomized Clinical Trial. JAMA Intern Med. 2018;178(11):1474-1481. 48 53. Vanelderen P, Van Zundert J, Kozicz T, et al. Effect of minocycline on lumbar radicular 49 neuropathic pain: a randomized, placebo-controlled, double-blind clinical trial with 50 amitriptyline as a comparator. Anesthesiology. 2015;122(2):399-406. 51 54. Effect of Milnacipran in Chronic Neuropathic Low Back Pain (NCT01225068). 2010. 52 55. Atkinson JH, Slater MA, Capparelli EV, et al. Efficacy of noradrenergic and serotonergic 53 antidepressants in chronic back pain: a preliminary concentration-controlled trial. J Clin 54 55 Psychopharmacol. 2007;27(2):135-142. 56 56. Atkinson JH, Slater MA, Wahlgren DR, et al. Effects of noradrenergic and serotonergic 57 antidepressants on chronic low back pain intensity. Pain. 1999;83(2):137-145. 58 57. Atkinson JH, Slater MA, Williams RA, et al. A placebo-controlled randomized clinical trial of 59 nortriptyline for chronic low back pain. Pain. 1998;76(3):287-296. 60

13 https://mc.manuscriptcentral.com/bmj BMJ Page 14 of 56

1 2 58. Hameroff SR, Cork RC, Weiss JL, Crago BR, Davis TP. Doxepin Effects on Chronic Pain and 3 Depression - a Controlled-Study. Adv Pain Res Ther. 1985;9:761-771. 4 59. Alcoff J, Jones E, Newman R. Controlled Trial of Imipramine for Chronic Low-Back-Pain. J Fam 5 Practice. 1982;14(5):841-846. 6 60. Hameroff SR, Cork RC, Scherer K, et al. Doxepin Effects on Chronic Pain, Depression and 7 8 Plasma Opioids. J Clin Psychiat. 1982;43(8):22-27. 9 61. Tetreault P, Mansour A, Vachon-Presseau E, Schnitzer TJ, Apkarian AV, Baliki MN. Brain 10 Connectivity Predicts Placebo Response across Chronic Pain Clinical Trials. Plos Biol. 11 2016;14(10). 12 62. Abou-RayaConfidential: S, Abou-Raya A, Helmii M.For Duloxetine Review for the management Only of pain in older adults 13 with knee osteoarthritis: randomised placebo-controlled trial (vol 41, pg 646, 2012). Age 14 Ageing. 2017;46(2):338-338. 15 63. Milnacipran for Chronic Pain in Knee Osteoarthritis (KOA) (NCT01510457). 2012. 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

14 https://mc.manuscriptcentral.com/bmj Page 15 of 56 BMJ

1 2 3 Figure legends 4 5 Figure 1. PRISMA Flow diagram 6 7 Figure 2. Summary of findings and certainty of evidence for pain outcomes. 8 9 Figure 3. Summary of findings and certainty of evidence for disability outcomes. 10 Figure 4. Summary of findings and certainty of evidence for safety outcomes. 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

15 https://mc.manuscriptcentral.com/bmj BMJ Page 16 of 56

1 2 3 Figure 2. Summary of findings and certainty of evidence for pain outcomes. 4 No. Certainty Antidepressant Mean difference Mean difference 5 Condition participants of evidence class (95 % CI), I2 (95% CI) 6 (trials) (GRADE) 7 Immediate term 8 9 SNRI 1,068 (3) -3.7 (-5.9 to -1.5), 0% Moderatea 10 Spinal pain SSRI 92 (1) -4.3 (-14.0 to 5.4), 0% Lowa,d 11 a,d 12 Confidential:TCA 145 (3) -0.9For (-5.4 to 3.6), Review 45% Only Low 13 14 SNRI 50 (1) -18.6 (-31.6 to -5.3), N/A Very lowa,c,d Sciatica 15 TCA 94 (2) -7.6 (-18.2 to 3.2), 0% Very lowa,c,d 16 17 18 Osteoarthritis SNRI 1,328 (4) -4.7 (-6.3 to -3.1), 0% Moderatea 19 Short term 20 a 21 SNRI 1,415 (4) -5.3 (-7.3 to -3.3), 0% Moderate 22 SSRI 170 (3) 1.5 (-5.4 to 8.4), 0% Lowa,d 23 TCA 591 (7) -9.4 (-21.5 to 1.6), 92% Very lowa,b.d 24 Spinal pain 25 NDRI 44 (1) -1.0 (-12.2 to 10.2), N/A Lowa,d 26 SARI 40 (1) -5.4 (-22.9 to 12.1), N/A Lowa,d 27 28 Tetracyclic 34 (1) -4.5 (-20.4 to 11.4), N/A 29 30 SNRI 96 (3) -17.5 (-42.9 to 7.9), 77% Very lowa,c,d 31 Sciatica TCA 116 (2) -16.0 (-31.5 to -0.4), 71% Very lowa,c,d 32 33 34 Osteoarthritis SNRI 1,941 (8) -9.2 (-11.6 to -6.8), 54% Lowa,b 35 Intermediate term 36 a,d 37 Spinal pain TCA 118 (1) -7.8 (-15.5 to 0.1), N/A Low 38 Sciatica TCA 60 (1) -27.0 (-36.1 to -17.9), N/A Lowa,d

39 -30 0 30 40 Favors antidepressants Favors placebo 41 42serotonin-norepinephrine reuptake inhibitor (SNRI), tricyclic antidepressants (TCA), serotonin reuptake inhibitor (SSRI), norepinephrine-dopamine reuptake inhibitor (NDRI), serotonin antagonist and reuptake inhibitors (SARI), tetracyclic antidepressants (tetracyclic); N/A, Not applicable 43aDowngraded by one level because > 25% of participants in this comparison are from studies at high risk of bias 44bDowngraded by one level because heterogeneity (I2) > 50% 45cDowngraded by one level because the limits of the 95% confidence interval are 20 points different to the point estimate dDowngraded by one level due to small study bias 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj Page 17 of 56 BMJ

1 2 3 Figure 3. Summary of findings and certainty of evidence for disability outcomes. 4 No. Certainty Antidepressant Mean difference Mean difference 5 Condition participants of evidence class (95 % CI), I2 (95% CI) 6 (trials) (GRADE) 7 Immediate term 8 9 Sciatica TCA 60 (1) -5.0 (-12.2 to 2.2), N/A Lowa,d 10 Osteoarthritis SNRI 353 (1) -5.1 (-7.3 to -2.9), N/A Moderatea 11 12 Short term Confidential: For Review Only 13 SNRI 1,423 (4) -3.6 (-5.2 to -1.9), 0% Moderatea 14 15 SSRI 92 (1) -2.2 (-8.1 to 3.7), N/A Lowa,d Spinal pain 16 TCA 398 (3) -12.2 (-25.6 to 1.3), 96% Very lowa,b,d 17 18 SARI 40 (1) 2.6 (-6.8 to 12.0), N/A Lowa,d 19 20 SNRI 8 (1) -4.4 (-19.9 to 11.1), N/A Very lowa,c,d Sciatica 21 TCA 116 (2) -8.4 (-18.2 to 1.4), 70% Very lowa,b,d 22 23 a,b 24 Osteoarthritis SNRI 1,810 (7) -6.0 (-8.1 to -4.0), 53% Low 25 Intermediate term 26 27 Spinal pain TCA 118 (1) -4.3 (-10.5 to 1.9), N/A Lowa,d 28 Sciatica TCA 60 (1) -20.0 (-27.7 to -12.3), N/A Lowa,d 29 -30 0 30 30 Favors antidepressants Favors placebo 31 32serotonin-norepinephrine reuptake inhibitor (SNRI), tricyclic antidepressants (TCA), serotonin reuptake inhibitor (SSRI), norepinephrine-dopamine reuptake inhibitor 33(NDRI), serotonin antagonist and reuptake inhibitors (SARI), tetracyclic antidepressants (tetracyclic); N/A, Not applicable a 34Downgraded by one level because > 25% of participants in this comparison are from studies at high risk of bias bDowngraded by one level because heterogeneity (I2) > 50% 35c Downgraded by one level because the limits of the 95% confidence interval are 20 points different to the point estimate 36dDowngraded by one level due to small study bias 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 18 of 56

1 2 3 Figure 4. Summary of findings and certainty of evidence for safety outcomes. 4 Certainty Antidepressant class No. events (%) Relative risk Relative risk 5 of evidence (trials) antidepressants / placebo (95 % CI), I2 (95% CI) 6 (GRADE) 7 Any adverse event 8 9 SNRI (13) 1,136 (62.7)/1,810 (49.8) 1.26 (1.19 to 1.34), 0% Lowa,c 10 TCA (7) 86 (29.9)/52 (19.5) 1.53 (1.13 to 2.1), 68.5% Very lowa,b,c,d 11 a,b,c,d 12 SSRI (2) Confidential:72 (67.9)/34 (63.0) 1.1 (0.85For to 1.38), 51.8%Review Only Very low 13 Tetracyclic (1) 18(90.0)/15 (96.9) 0.96 (0.79 to 1.16), N/A Lowa,d 14 Serious adverse events 15 16 SNRI (10) 29 (1.7)/20 (1.3) 1.31 (0.74 to 2.31), 0% Very lowa,c,d 17 TCA (1) 1 (2.6)/0 (0) 2.62 (0.11 to 62.1), N/A Very lowa,c,d 18 19 NDRI (1) 2 (4.5)/1 (2.3) 2.0 (0.19 to 21.26), N/A Lowa,c 20 SARI (1) 3 (13.6)/1 (5.0) 2.73 (0.31 to 24.14), N/A Very lowa,c,d 21 22 Drop-outs due to adverse events 23 SNRI (11) 230 (12.3)/89 (5.2) 2.35 (1.85 to 2.97), 0% Very lowa,c,d 24 a,c,d 25 TCA (12) 56 (11.5)/22 (5.1) 2.32 (1.44 to 3.75), 0% Very low 26 NDRI (1) 3 (5.6)/3 (5.6) 1.0 (0.21 to 4.74), N/A Lowa,c 27 SARI (1) 3 (13.6)/1 (5.0) 2.73 (0.31 to 24.14), N/A Very lowa,c,d 28 0.1 0 10 29 Favors antidepressants Favors placebo 30 31serotonin-norepinephrine reuptake inhibitor (SNRI), tricyclic antidepressants (TCA), serotonin reuptake inhibitor (SSRI), norepinephrine-dopamine reuptake inhibitor (NDRI), serotonin antagonist and reuptake inhibitors (SARI), tetracyclic antidepressants (tetracyclic); 32N/A, Not applicable 33aDowngraded by one level because > 25% of participants in this comparison are from studies at high risk of bias 34bDowngraded by one level because heterogeneity (I2) > 50% c 35Downgraded by one level because the 95% CI include appreciable benefit (i.e. 95% CI under 0.75) or harm (i.e. 95% CI over 1.25) dDowngraded by one level due to small study bias 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 Page 19 of 56 BMJ

1 2 3 Figure 1. PRISMA Flow diagram 4 5 Records identified through database searching (n = 2,551) 6 - Cochrane CENTRAL (n = 218) 7 - CINAHL (n = 58) 8 - Embase (n = 1,969) 9 - International Pharmaceutical Abstracts (n = 28) 10 - Medline (n = 278) 11 Confidential: For Review Only 12 Records identified through clinical trials registries (n = 220) 13 - ClinicalTrials (n = 67) - World Health Organization Clinical Trials Registry Platform (ICTRP) (n = 153) 14 15 16 17 18 19 20 Records after duplicates removed Records excluded 21 (n = 1,930) (n = 1,795) 22 23 24 25 26 27 Full-text articles assessed for Full-text articles excluded, with 28 reasons (n = 102) 29 eligibility (n = 135) 30 - Completed, no results (n = 8) 31 - Conference abstract (n = 12) 32 - Invalid results (n = 1) 33 Studies included in the descriptive - Combined data from multiple 34 RCTs (n = 3) analysis (n = 33) 35 - Recruiting (n = 5) 36 - Terminated/unknown/withdrawn 37 (n = 5) 38 - Wrong comparator (n = 18) 39 - Wrong intervention (n = 17) 40 Studies included in the - Wrong outcomes (n = 2) 41 meta-analysis (n = 33) - Wrong population (n = 13) 42 - Wrong study design (18) 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 20 of 56

1 2 3 4 Efficacy and safety of antidepressants for the treatment of spinal 5 6 pain and osteoarthritis: A systematic review and meta-analysis 7 8 9 10 11 12 Confidential: For Review Only 13 14 15 16 Table of contents 17 18 19 Supplemental file 1. Search strategies ...... 2 20 21 Supplemental file 2. List of excluded studies after full-text reading...... 8 22 23 Supplemental file 3. Calculation of effect sizes for immediate, short term and intermediate term pain in 24 spinal pain and sciatica trials ...... 16 25 26 Supplemental file 4. Calculation of effect sizes for immediate, short and intermediate term pain in 27 28 osteoarthritis trials...... 20 29 30 Supplemental file 5. Calculation of effect sizes for immediate, short term and intermediate term 31 32 disability in spinal pain trials...... 22 33 34 Supplemental file 6. Calculation of effect sizes for immediate, short term and intermediate term 35 disability in osteoarthritis trials...... 24 36 37 Supplemental file 7. Number of serious adverse events, adverse events, and drop-outs due to adverse 38 39 events and in spinal pain and osteoarthritis trials...... 26 40 41 Supplemental file 8. GRADE Framework...... 29 42 Supplemental file 9. Characteristics of the included studies (n = 33) ...... 30 43 44 Control: Active placebo (Benztropine 0.5 mg/day)...... 32 45 46 Rescue medication: Not mentioned ...... 33 47 48 Supplemental file 10. Risk of bias of included trials...... 37 49 50 Supplemental file 11. Exploratory meta-regression ...... 38 51 52 53 54 55 56 57 58 59 60

1 https://mc.manuscriptcentral.com/bmj Page 21 of 56 BMJ

1 2 3 Supplemental file 1. Search strategies 4 5 MEDLINE 6 7 1 antidepressants.mp. or exp Antidepressive Agents/ 8 2 exp Antidepressive Agents, Second-Generation/ 9 3 exp Antidepressive Agents, Tricyclic/ 10 11 4 exp Amitriptyline/ 12 5 Confidential:exp Nortriptyline/ For Review Only 13 6 exp Desipramine/ 14 7 exp Imipramine/ 15 8 exp Doxepin/ 16 9 exp Trimipramine/ 17 10 exp Clomipramine/ 18 11 exp Protriptyline/ 19 12 Tetracyclic.mp. 20 13 exp Amoxapine/ 21 22 14 exp Maprotiline/ 23 15 exp Serotonin Uptake Inhibitors/ 24 16 exp Citalopram/ 25 17 Escitalopram.mp.Fluoxetine 26 18 exp Fluvoxamine/ 27 19 exp Paroxetine/ 28 20 exp Sertraline/ 29 21 Duloxetine.mp. or exp Duloxetine Hydrochloride/ 30 22 Venlafaxine.mp. or exp Venlafaxine Hydrochloride/ 31 23 Desvenlafaxine.mp. or exp Desvenlafaxine Succinate/ 32 24 exp Milnacipran/ 33 34 25 exp Monoamine Oxidase Inhibitors/ 35 26 exp Harmaline/ 36 27 exp Iproniazid/ 37 28 exp Isocarboxazid/ 38 29 exp Moclobemide/ 39 30 exp Tranylcypromine/ 40 31 exp Bupropion/ 41 32 exp Trazodone/ 42 33 exp Mirtazapine/ 43 34 1 OR 2 OR 3 OR 4 OR 5 OR 6 OR 7 OR 8 OR 9 OR 10 OR 11 OR 12 OR 13 44 OR 14 OR 15 OR 16 OR 17 OR 18 OR 19 OR 20 OR 21 OR 22 OR 23 OR 24 45 46 OR 25 OR 26 OR 27 OR 28 OR 29 OR 30 OR 31 OR 32 OR 33 47 35 osteoarthritis.mp. OR exp osteoarthritis/ 48 36 exp low back pain/ 49 37 exp back pain/ 50 38 exp neck pain/ 51 39 ("low back pain" OR "back pain" OR "neck pain" OR backache OR lumbago 52 OR "neck ache" OR "spin* pain" OR "knee pain" OR "hip pain").mp. 53 40 exp Sciatica/ 54 41 exp Radiculopathy/ 55 42 35 OR 36 OR 37 OR 38 OR 39 OR 40 OR 41 56 43 34 AND 42 57 58 44 randomized controlled trial.pt. OR exp randomized controlled trial/ 59 45 "randomized controlled trial".mp. 60 46 exp random allocation/

2 https://mc.manuscriptcentral.com/bmj BMJ Page 22 of 56

1 2 3 47 placebo.mp. OR exp placebos/ OR exp placebo effect/ 4 48 (random* adj3 trial).ab,ti. 5 49 "controlled clinical trial".mp. OR exp controlled clinical trial/ 6 7 50 Random*.ab,ti. 8 51 44 OR 45 OR 46 OR 47 OR 48 OR 49 OR 50 9 52 43 AND 51 10 11 12 Pubmed Confidential: For Review Only 13 (((((((Low Back Pain OR back pain OR neck pain OR Sciatica OR Radiculopathy[MeSH Terms])) OR 14 15 (radiculopathies[Title/Abstract] OR radiculopathy, cervical[Title/Abstract] OR Cervical 16 Radiculopathies[Title/Abstract] OR Cervical Radiculopathy[Title/Abstract] OR Radiculopathies, 17 Cervical[Title/Abstract] OR Nerve Root Disorder[Title/Abstract] OR Radiculitis[Title/Abstract] OR 18 Nerve Root Inflammation[Title/Abstract] OR Compression, Nerve Root[Title/Abstract] OR 19 lumbago[Title/Abstract] OR lower back pain[Title/Abstract] OR low back ache[Title/Abstract] OR low 20 backaches[Title/Abstract] OR postural low back pain[Title/Abstract] OR mechanical low back 21 pain[Title/Abstract] OR recurrent low back pain[Title/Abstract] OR vertebrogenic pain 22 syndrome[Title/Abstract] OR back aches[Title/Abstract] OR back pain with radiation[Title/Abstract] 23 OR neckache[Title/Abstract] OR cervicalgia[Title/Abstract] OR cervicodynias[Title/Abstract] OR 24 cervical pain[Title/Abstract] OR posterior cervical pain[Title/Abstract] OR anterior neck 25 pain[Title/Abstract])) OR (Osteoarthritis OR Osteoarthritis, Spine OR Osteoarthritis, Knee OR 26 Osteoarthritis OR Osteoarthritis, Spine OR Osteoarthritis, Knee[MeSH Terms])) OR 27 (Osteoarthrosis[Title/Abstract] OR Osteoarthroses[Title/Abstract] OR Osteoarthritides[Title/Abstract] 28 OR Arthritis, Degenerative[Title/Abstract] OR Arthritides, Degenerative[Title/Abstract] OR 29 Degenerative Arthritides[Title/Abstract] OR Degenerative Arthritis[Title/Abstract] OR Osteoarthrosis 30 Deformans[Title/Abstract] OR Coxarthrosis[Title/Abstract] OR Coxarthroses[Title/Abstract] OR 31 Osteoarthritis Of Hip[Title/Abstract] OR Osteoarthritis Of Hips[Title/Abstract] OR Knee 32 Osteoarthritides[Title/Abstract] OR Osteoarthritides, Knee[Title/Abstract] OR Osteoarthritis Of 33 Knees[Title/Abstract]))) 34 35 AND 36 (((Randomized Controlled Trial [Publication Type] OR Randomized Controlled Trials as Topic OR 37 Controlled Clinical Trial [Publication Type] OR quivalence Trial [Publication Type] OR Random 38 Allocation OR Cross-Over Studies OR Double-Blind Method OR Placebos)) OR 39 (randomization[Title/Abstract] OR Cross-Over Study[Title/Abstract] OR Crossover 40 Studies[Title/Abstract] OR Cross-Over Studies[Title/Abstract] OR Crossover Study[Title/Abstract] 41 OR Crossover Trials[Title/Abstract] OR Crossover design[Title/Abstract] OR Double-Blind 42 Study[Title/Abstract] OR Double-Blind Studies[Title/Abstract] OR allocation[Title/Abstract] OR 43 random[Title/Abstract] OR follow-up[Title/Abstract] OR active placebo[Title/Abstract] OR 44 placebo[Title/Abstract] OR non-inferiority[Title/Abstract] OR superiority[Title/Abstract]))) 45 46 AND 47 ((((Antidepressive Agents OR Antidepressive Agents, Tricyclic OR Antidepressive Agents, Second- 48 Generation OR Antidepressive Agents, Second-Generation [Pharmacological Action] OR 49 Antidepressive Agents, Tricyclic [Pharmacological Action] OR Adrenergic Uptake Inhibitors OR 50 Antidepressive Agents OR Antidepressive Agents, Tricyclic OR Antidepressive Agents, Second- 51 Generation OR Antidepressive Agents, Second-Generation [Pharmacological Action] OR 52 Antidepressive Agents, Tricyclic [Pharmacological Action] OR Adrenergic Uptake Inhibitors OR))) 53 54 OR ((Amitriptyline OR Nortriptyline OR Desipramine OR Imipramine OR Doxepin OR Trimipramine 55 OR Clomipramine OR Protriptyline OR Amoxapine OR Maprotiline OR Serotonin Uptake Inhibitors 56 OR Citalopram OR Fluoxetine OR Paroxetine OR Sertraline OR Duloxetine Hydrochloride OR 57 Fluvoxamine OR Venlafaxine Hydrochloride OR Desvenlafaxine Succinate OR Milnacipran OR 58 Monoamine Oxidase Inhibitors OR Harmaline OR Iproniazid OR Isocarboxazid OR Moclobemide OR 59 Tranylcypromine OR Bupropion OR Trazodone OR Mirtazapine[MeSH Terms]))) 60

3 https://mc.manuscriptcentral.com/bmj Page 23 of 56 BMJ

1 2 3 EMBASE 4 5 1 exp knee osteoarthritis/ or exp hip osteoarthritis/ or exp osteoarthritis/ 6 2 exp low back pain/ 7 3 exp neck pain/ 8 4 exp backache/ or exp intervertebral disk hernia/ or exp sciatica/ 9 5 exp radiculopathy/ 10 11 6 exp cervicobrachial neuralgia/ 12 7 Confidential:random:.tw. For Review Only 13 8 placebo:.mp. 14 9 double-blind:.tw. 15 10 antidepressants.mp. or exp antidepressant agent/ 16 11 exp amitriptyline/ 17 12 exp nortriptyline/ 18 13 exp desipramine/ 19 14 Exp imipramine/ 20 15 exp doxepin/ 21 16 exp trimipramine/ 22 23 17 exp clomipramine/ 24 18 exp protriptyline/ 25 19 exp tetracyclic antidepressant agent/ 26 20 exp amoxapine/ 27 21 exp maprotiline/ 28 22 exp serotonin uptake inhibitor/ 29 23 exp citalopram/ 30 24 exp fluoxetine/ 31 25 exp fluvoxamine/ 32 26 exp paroxetine/ 33 34 27 exp sertraline/ 35 28 exp duloxetine/ 36 29 exp venlafaxine/ 37 30 exp desvenlafaxine/ 38 31 exp milnacipran/ 39 32 exp monoamine oxidase inhibitor/ 40 33 exp harmaline/ 41 34 exp iproniazid/ 42 35 exp isocarboxazid/ 43 36 exp moclobemide/ 44 37 exp tranylcypromine/ 45 46 38 Bupropion.mp. or exp amfebutamone/ 47 39 exp trazodone/ 48 40 exp mirtazapine/ 49 41 1 or 2 or 3 or 4 or 5 or 6 50 42 7 or 8 or 9 51 43 10 or 11 or 12 or 13 or 14 or 15 or 16 or 17 or 18 or 19 or 20 or 21 or 22 or 23 or 24 or 52 25 or 26 or 27 or 28 or 29 or 30 or 31 or 32 or 33 or 34 or 35 or 36 or 37 or 38 or 39 or 53 40 54 44 41 and 42 and 43 55 56 57 58 59 60

4 https://mc.manuscriptcentral.com/bmj BMJ Page 24 of 56

1 2 3 Cochrane Central 4 5 Low Back Pain OR back pain OR neck pain OR sciatica OR Radiculopathy OR Osteoarthritis OR 6 Osteoarthritis, Spine OR Osteoarthritis, Knee OR Osteoarthritis, Hip 7 AND 8 Antidepressive drugs OR antidepressive agents OR Antidepressive Agents, Tricyclic OR 9 Amitriptyline OR Nortriptyline OR Desipramine OR Imipramine OR Doxepin OR Trimipramine OR 10 Clomipramine OR Protriptyline OR Amoxapine OR Maprotiline OR Serotonin Uptake Inhibitors OR 11 Citalopram OR Fluoxetine OR Paroxetine OR Sertraline OR Duloxetine Hydrochloride OR 12 Confidential: For Review Only 13 Fluvoxamine OR Venlafaxine Hydrochloride OR Desvenlafaxine Succinate OR Milnacipran OR 14 Monoamine Oxidase Inhibitors OR Harmaline OR Iproniazid OR Isocarboxazid OR Moclobemide 15 OR Tranylcypromine OR Bupropion OR Trazodone OR Mirtazapine 16 AND 17 Randomized Controlled Trial OR Double-Blind Method OR Placebos OR Random Allocation 18 19 20 International Pharmaceutical Abstracts 21 22 1 osteoarthritis.mp. 23 2 knee osteoarthritis.mp. 24 3 hip osteoarthritis.mp. 25 4 neck pain.mp. 26 5 low back pain.mp. 27 6 back pain.mp. 28 7 backache.mp. 29 8 sciatica.mp. 30 9 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 31 10 random.mp. 32 33 11 placebo.mp. 34 12 double-blind.mp. 35 13 10 or 11 or 12 36 14 antidepressants.mp 37 15 amitriptyline.mp. 38 16 nortriptyline.mp 39 17 desipramine.mp. 40 18 imipramine.mp. 41 19 doxepin.mp. 42 20 trimipramine.mp. 43 21 clomipramine.mp. 44 45 22 protriptyline.mp. 46 23 tetracyclic antidepressant agent.mp. 47 24 amoxapine.mp. 48 25 maprotiline.mp. 49 26 serotonin uptake inhibitor.mp. 50 27 citalopram.mp. 51 28 fluoxetine.mp. 52 29 fluvoxamine.mp. 53 30 paroxetine.mp. 54 31 sertraline.mp. 55 56 32 duloxetine.mp. 57 33 venlafaxine.mp. 58 34 desvenlafaxine.mp. 59 35 milnacipran.mp 60 36 monoamine oxidase inhibitor.mp.

5 https://mc.manuscriptcentral.com/bmj Page 25 of 56 BMJ

1 2 3 37 harmaline.mp. 4 38 iproniazid.mp. 5 39 isocarboxazid.mp. 6 7 40 moclobemide.mp. 8 41 tranylcypromine.mp. 9 42 Bupropion.mp. 10 43 trazodone.mp. 11 44 mirtazapine.mp. 12 45 14 Confidential:OR 15 OR 16 OR 17 OR 18 OR 19 For OR 20 OR Review 21 OR 22 OR 23 OROnly 24 OR 25 OR 26 OR 27 13 OR 28 OR 29 OR 30 OR 31 OR 32 OR 33 OR 34 OR 35 OR 36 OR 37 OR 38 OR 39 OR 40 14 OR 41 OR 42 OR 43 OR 44 15 46 9 AND 13 AND 45 16 17 18 CINAHL 19 20 1 (MH "Osteoarthritis+") OR (MH "Osteoarthritis, Spine+") OR (MH "Osteoarthritis, Knee") 21 OR (MH "Osteoarthritis, Hip") OR (MH "Osteoarthritis, Cervical") 22 2 (MH "Low Back Pain") OR (MH "Back Pain+") OR (MH "Neck Pain") 23 3 (MH "Sciatica") 24 4 (MH "Radiculopathy") 25 5 S1 OR S2 OR S3 OR S4 26 6 (MH "Randomized Controlled Trials+") OR (MH "Clinical Trials+") 27 28 7 (MH "Placebos") 29 8 (MH "Double-Blind Studies") 30 9 (MH "Random Sample+") OR (MH "Random Assignment") 31 10 S6 OR S7 OR S8 OR S9 32 11 (MH "Antidepressive Agents+") 33 12 (MH "Amitriptyline") 34 13 (MH "Nortriptyline") 35 14 (MH "Desipramine") 36 15 (MH "Imipramine") 37 16 (MH "Doxepin") 38 39 17 "trimipramine" 40 18 (MH "Clomipramine") 41 19 "protriptyline" 42 20 (MH "Amoxapine") 43 21 (MH "Maprotiline") 44 22 (MH "Citalopram") 45 23 (MH "Fluoxetine+") 46 24 (MH "Fluvoxamine Maleate") 47 25 (MH "Paroxetine") 48 26 (MH "Sertraline Hydrochloride") 49 50 27 (MH "Duloxetine Hydrochloride") 51 28 (MH "Venlafaxine+") 52 29 (MH "Desvenlafaxine Succinate") 53 30 (MH "Milnacipran Hydrochloride") 54 31 "harmaline" 55 32 "iproniazid" 56 33 "isocarboxazid" 57 34 "moclobemide" 58 35 "tranylcypromine" 59 36 (MH "Bupropion") 60

6 https://mc.manuscriptcentral.com/bmj BMJ Page 26 of 56

1 2 3 37 (MH "Trazodone") 4 38 (MH "Mirtazapine") 5 39 S11 OR S12 OR S13 OR S14 OR S15 OR S16 OR S17 OR S18 OR S19 OR S20 OR S21 6 OR S22 OR S23 OR S24 OR S25 OR S26 OR S27 OR S28 OR S29 OR S30 OR S31 OR S32 7 8 OR S33 OR S34 OR S35 OR S36 OR S37 OR S38 9 40 S5 AND S10 AND S39 10 11 12 ClinicalTrials.govConfidential: and International Clinical ForTrials Registry Review Platform* Only 13 *Each antidepressant was searched individually 14 15 1. Low back pain AND (Amitriptyline OR Nortriptyline OR Desipramine OR Imipramine OR 16 Doxepin OR Trimipramine OR Clomipramine OR protriptyline OR Amoxapine OR Maprotiline OR 17 serotonin uptake inhibitor OR Citalopram OR fluoxetine OR Fluvoxamine OR Paroxetine Sertraline 18 OR Duloxetine OR Venlafaxine OR Desvenlafaxine OR Milnacipran OR Harmaline OR Iproniazid 19 OR Isocarboxazid OR Moclobemide OR Tranylcypromine OR Bupropion OR Trazodone OR 20 mirtazapine) 21 22 2. Neck pain AND (Amitriptyline OR Nortriptyline OR Desipramine OR Imipramine OR Doxepin 23 OR Trimipramine OR Clomipramine OR protriptyline OR Amoxapine OR Maprotiline OR serotonin 24 uptake inhibitor OR Citalopram OR fluoxetine OR Fluvoxamine OR Paroxetine Sertraline OR 25 Duloxetine OR Venlafaxine OR Desvenlafaxine OR Milnacipran OR Harmaline OR Iproniazid OR 26 Isocarboxazid OR Moclobemide OR Tranylcypromine OR Bupropion OR Trazodone OR 27 mirtazapine) 28 3. Sciatica AND (Amitriptyline OR Nortriptyline OR Desipramine OR Imipramine OR Doxepin OR 29 Trimipramine OR Clomipramine OR protriptyline OR Amoxapine OR Maprotiline OR serotonin 30 uptake inhibitor OR Citalopram OR fluoxetine OR Fluvoxamine OR Paroxetine Sertraline OR 31 Duloxetine OR Venlafaxine OR Desvenlafaxine OR Milnacipran OR Harmaline OR Iproniazid OR 32 Isocarboxazid OR Moclobemide OR Tranylcypromine OR Bupropion OR Trazodone OR 33 mirtazapine) 34 4. Osteoarthritis AND (Amitriptyline OR Nortriptyline OR Desipramine OR Imipramine OR 35 Doxepin OR Trimipramine OR Clomipramine OR protriptyline OR Amoxapine OR Maprotiline OR 36 serotonin uptake inhibitor OR Citalopram OR fluoxetine OR Fluvoxamine OR Paroxetine Sertraline 37 OR Duloxetine OR Venlafaxine OR Desvenlafaxine OR Milnacipran OR Harmaline OR Iproniazid 38 39 OR Isocarboxazid OR Moclobemide OR Tranylcypromine OR Bupropion OR Trazodone OR 40 mirtazapine) 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

7 https://mc.manuscriptcentral.com/bmj Page 27 of 56 BMJ

1 2 3 Supplemental file 2. List of excluded studies after full-text reading 4 Author/registry ID Year Title Reason for exclusion 5 identification 6 An Open-Label, 52-Week, Phase III Trial of Duloxetine in Japanese Patients with 7 1 Konno 2019 Wrong study design 8 Chronic Low Back Pain Confidential:Clinical meaningfulness of duloxetine's For effect Review in Chinese patients with chronicOnly pain 9 2 Yue 2019 Wrong study design 10 due to osteoarthritis: post hoc analyses of a phase 3 randomized trial Efficacy of duloxetine and gabapentin in pain reduction in patients with knee 11 3 Enteshari-Moghaddam 2019 Wrong comparator 12 osteoarthritis 13 Efficacy of duloxetine for multi-site pain in patients with knee pain due to 4 Itoh 2019 Conference abstract 14 osteoarthritis: an exploratory post hoc analysis of a Japanese phase 3 randomized study 15 Maintenance of effect of duloxetine in Chinese patients with pain due to osteoarthritis: 5 Wang 2019 Wrong study design 16 13-week open-label extension data 17 Opioid Exposure Negatively Affects Antidepressant Response to Venlafaxine in Older 6 Stahl 2019 Wrong study design 18 Adults with Chronic Low Back Pain and Depression 19 Efficacy of duloxetine by prior NSAID use in the treatment of chronic osteoarthritis 20 7 Enomoto 2018 knee pain: A post hoc subgroup analysis of a randomized, placebo-controlled, phase 3 Wrong study design 21 study in Japan 22 Response to duloxetine in patients with knee pain due to osteoarthritis: an exploratory 23 8 Itoh 2018 Wrong study design 24 post hoc analysis of a Japanese Phase III randomized study Safety and efficacy of duloxetine in Japanese patients with chronic knee pain due to 25 9 Uchio 2018 Wrong study design 26 osteoarthritis: an open-label, long-term, Phase III extension study Duloxetine 60 mg for chronic low back pain: post hoc responder analysis of double- Combined data from 27 10 Alev 2017 28 blind, placebo-controlled trials other RCTs 29 Response to duloxetine in chronic low back pain: exploratory post hoc analysis of a 11 Tsuji 2017 Wrong study design 30 Japanese Phase III randomized study 31 Maintenance of effect and long-term safety of duloxetine in Chinese patients with pain 12 Wang 2016 Conference abstract 32 due to osteoarthritis 33 A randomized controlled trial of pregabalin versus amitriptyline in chronic low 13 Kalita 2014 Wrong comparator 34 backache 35 Efficacy of duloxetine versus placebo in patients with chronic low back pain and a 36 14 Schukro 2014 Conference abstract 37 neuropathic component 38 15 Gelijkens 2014 The effectiveness of amitriptyline in the treatment of subacute lumbar radicular pain Conference abstract 39 40 41 42 8 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 56

1 2 3 Efficacy of duloxetine versus alternative oral therapies: an indirect comparison of 16 Cawston 2013 Wrong study design 4 randomised clinical trials in chronic low back pain 5 Functional testing and pre-to post pain measure outcomes in chronic knee osteoarthritis 6 17 Mathur 2013 pain with milnacipran: Interim analysis of a randomized, double-blind, placebo- Conference abstract 7 8 controlled trial Confidential:Responsiveness of the Intermittent For and Constant Review Osteoarthritis Pain (ICOAP) Only scale in 9 18 Risser 2013 Wrong study design 10 a trial of duloxetine for treatment of osteoarthritis knee pain Safety and efficacy of duloxetine treatment in older and younger patients with 11 Combined data from 19 Micca 2013 osteoarthritis knee pain: a post hoc, subgroup analysis of two randomized, placebo- 12 other RCTs 13 controlled trials 14 Safety of duloxetine for the treatment of older patients with osteoarthritis knee pain or 20 Alaka 2013 Conference abstract 15 chronic low back pain 16 21 Weiss-Citrome 2012 Antidepressant Therapy for Pain Associated With Osteoarthritis Wrong study design 17 Clinically relevant outcomes based on analysis of pooled data from 2 trials of Combined data from 18 22 Hochberg 2012 19 duloxetine in patients with knee osteoarthritis other RCTs 20 23 Ivanova 2012 Duloxetine use in chronic low back pain: treatment patterns and costs Wrong study design 21 22 Duloxetine as treatment for knee pain in patients with osteoarthritis who regularly use 23 24 Hochberg 2011 nonsteroidal anti-inflammatory drugs (NSAIDS): A post hoc analysis of two Conference abstract 24 randomized, placebo-controlled trials 25 25 Johnson 2011 Effects of duloxetine and placebo in patients with chronic low back pain Conference abstract 26 The efficacy and safety of duloxetine treatment in older patients with osteoarthritis 27 26 Micca 2011 Conference abstract 28 knee pain: A post hoc, subgroup analysis of data from 2 placebo-controlled trials Duloxetine and care management treatment of older adults with comorbid major 29 27 Karp 2010 Wrong study design 30 depressive disorder and chronic low back pain: results of an open-label pilot study Efficacy and safety of duloxetine 60 mg once-daily in patients with chronic low back 31 28 Skljarevski 2010 Conference abstract 32 pain 33 Escitalopram 20 mg versus duloxetine 60 mg for the treatment of chronic low back 29 Mazza 2010 Wrong comparator 34 pain 35 Maintenance of effect of duloxetine in patients with chronic low back pain: a 41-week 30 Skljarevski 2010 Wrong study design 36 uncontrolled, dose-blinded study 37 A single-blind, placebo run-in study of duloxetine for activity-limiting osteoarthritis 31 Sullivan 2009 Wrong study design 38 pain 39 40 41 42 9 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 56 BMJ

1 2 3 Efficacy and safety of duloxetine in the treatment of pain associated with knee 32 Skljarevski 2009 Conference abstract 4 osteoarthritis 5 6 33 Skljarevski 2009 Maintenance of effect of duloxetine in patients with chronic low back pain Conference abstract 7 [Evaluation of analgesic action of fluvoxamine compared with efficacy of imipramine 34 Kwasucki 2002 Wrong comparator 8 Confidential:and tramadol for treatment of sciatica--openFor Reviewtrial] Only 9 The efficacy of amitriptyline and acetaminophen in the management of acute low back 35 Stein 1996 Wrong comparator 10 pain 11 Prospective study of the analgesic action of clomipramine versus placebo in refractory 12 36 Trèves 1991 Wrong outcomes lumbosciatica (68 cases) 13 The role of an antidepressant, dibenzepin (Noveril), in the relief of pain in chronic 14 37 Thorpe 1974 Wrong population 15 arthritic states Invalid results due to 16 38 NCT00945945 2009 A Study of Duloxetine in Patients With Osteoarthritis Knee Pain 17 technical problems Evaluation of Stepped Care for Chronic Pain in Iraq/Afghanistan Veterans (ESCAPE) 18 39 NCT00386243 2006 Wrong intervention 19 (ESCAPE) 20 40 NCT00009672 2001 Pain Treatment for Sciatica Completed, no results 21 22 41 NCT00227292 2005 Cipralex in Treatment of Depressive Symptoms and Chronic Back Pain Withdrawn 23 Comparison of the Efficacy of Duloxetine With Placebo in Patients With Chronic Low 24 42 NCT01166048 2010 Completed, no results Back Pain With a Radicular Component 25 26 43 NCT01221740 2010 Pain Assessment and Quality of Life in Back Pain Patients: Role of Milnacipran Withdrawn 27 Study to Evaluate the Efficacy of Milnacipran in the Treatment of Pain Due to 44 NCT01329406 2011 Unknown 28 Osteoarthritis 29 30 45 NCT01377038 2011 OASIS: Osteoarthritis Sensitivity Integration Study (OASIS) Withdrawn 31 46 NCT01716377 2012 Healing With Venlafaxine After Injury (HELP) (HELP) Completed, no results 32 33 47 NCT02208778 2014 Imaging Pain Relief in Osteoarthritis (IPRO) Completed, no results 34 35 48 NCT03249558 2017 Effect of Combined Morphine and Duloxetine on Chronic Pain (Duloxetine) Recruiting 36 Genetic Variants Associated With Low Back Pain and Their Response to Treatment 49 NCT03364075 2017 Terminated 37 With Duloxetine or Propranolol 38 ACTRN1261400068363 50 2014 A randomised placebo controlled trial of nortriptyline for pain in knee osteoarthritis Completed, no results 39 9 40 41 42 10 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 56

1 2 3 Does low dose amitriptyline reduce pain in knee osteoarthritis? A double blind, ACTRN1261500030156 4 51 2015 randomised, pragmatic, placebo controlled clinical trial of amitriptyline compared to Recruiting 1 5 active placebo in addition to usual care 6 ACTRN1261900087817 A randomised controlled trial of venlafaxine to treat patients with knee osteoarthritis 7 52 2019 Not yet recruiting 8 8 pain ACTRN1261900108219Confidential:Venlafaxine compared to duloxetine For for the treatmentReview of osteoarthritis pain: Only A double- 9 53 2019 Recruiting 10 0 blind, randomised, non-inferiority trial 11 54 JPRN-JapicCTI-132129 2013 A phase 3 clinical trial of duloxetine in patients with chronic low back pain Completed, no results 12 13 55 JPRN-JapicCTI-142676 2014 A phase 3 clinical trial of duloxetine in patients with osteoarthritis Completed, no results 14 EUCTR2009-012713- Randomized Double-blind Study Comparing the Efficacy of Duloxetine with Placebo 56 2009 Completed, no results 15 22-AT in Patients with Chronic Low Back Pain 16 A mechanism based proof of concept study of the effects of duloxetine in the treatment 57 NCT04224584 2020 Recruiting 17 of patients with osteoarthritic knee pain 18 Effectiveness and cost-effectiveness of duloxetine added to usual care for patients with 19 58 NTR4798 2017 chronic pain due to hip or knee osteoarthritis: protocol of a pragmatic open-label cluster Wrong comparator 20 21 randomised trial (the DUO trial). A Single-Blind Placebo Run-in Study of Duloxetine for Activity-Limiting 22 59 NCT00609557 2008 Wrong study design 23 Osteoarthritis Pai A Single-Blind Placebo Run-In Study of Venlafaxine for Activity-Limiting 24 60 NCT00611676 2008 Wrong study design 25 Osteoarthritis Pain 26 Duloxetine Treatment of Major Depression and Chronic Low Back Pain For Older 61 NCT00696293 2008 Wrong study design 27 Adults (ACHIEVE2) 28 Pre-emptive Effect of Duloxetine in the Second Knee in Staged Total Knee 62 NCT03792828 2019 Wrong comparator 29 Arthroplasty 30 31 63 NCT04111627 2019 Exercise Plus Duloxetine for Knee Osteoarthritis and Depression Wrong comparator 32 Duloxetine Combined With Intra-articular Injection of Corticosteroid and Hyaluronic 64 NCT04117893 2019 Wrong comparator 33 Acid Reduces Pain in the Treatment of Knee Osteoarthritis Patients 34 Efficacy and safety of pregabalin vs amitriptyline in chronic low back pain – A 65 CTRI/2012/07/002833 2012 Wrong comparator 35 randomized control trial 36 Study Comparing A New Drug Containing The Combination Meloxicam And 66 NCT01587508 2012 Wrong comparator 37 Cyclobenzaprine In The Treatment Of Acute Lumbago 38 An Extension Study of Duloxetine in Osteoarthritis and Knee Pain (Extension of F1J- 39 67 NCT02335346 2015 Wrong comparator JE-HMGX, NCT02248480) 40 41 42 11 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 56 BMJ

1 2 3 Effects of Duloxetine on Pain Relief After Total Knee Arthroplasty in Central 68 NCT02600247 2015 Wrong comparator 4 Sensitization Patient 5 An Open Label Extension Study of Duloxetine (LY248686) in Participants With 6 69 NCT01914666 2013 Wrong comparator Chronic Low Back Pain 7 IRCT20170716035126N Evaluating the efficacy of duloxetine and gabapentin in pain reduction in patients with 8 70 Confidential:2018 For Review OnlyWrong comparator 9 2 knee osteoarthritis 10 71 NCT01124188 2010 ADAPT: Addressing Depression and Pain Together (ADAPT) Wrong comparator 11 An open label extension study of phase 3 clinical trial of duloxetine in patients with 12 72 JPRN-JapicCTI-132312 2013 Wrong comparator chronic low back pain 13 A comparison of analgesic efficacy between Amitriptyline and Mianserin in chronic 14 73 TCTR20190303001 2018 Wrong comparator 15 low back pain patients: A randomized double-blind controlled trial 16 74 NCT01144923 2010 Conservative Therapy Versus Epidural Steroids for Cervical Radiculopathy Wrong intervention 17 Neurophysiological Basis of Rehabilitation in Complex Regional Pain Syndrome, 18 75 NCT02347579 2015 Wrong intervention Type I and Chronic Low Back Pain (BrainEXPain) 19 20 76 NCT02453802 2015 Comparison of Topical and Infusion Tranexamic Acid After Total Knee Arthroplasty Wrong intervention 21 The Blood Saving Effect of Tranexamic Acid in Total Knee Arthroplasty With 77 NCT02458729 2015 Wrong intervention 22 Rivaroxaban as Thromboprophylaxis 23 24 78 NCT02865174 2016 Topical Tranexamic Acid and Floseal® in Total Knee Arthroplasty Wrong intervention 25 Trial Evaluating multimOdal toPical Cream In CompArison to pLacebo (TOPICAL) 79 NCT03199417 2017 Wrong intervention 26 (TOPICAL) 27 28 80 NCT03328832 2017 Combined Topical Tranexamic Acid With Floseal® in Total Knee Arthroplasty Wrong intervention 29 ACTRN1261700112333 Safety and Efficacy of Autologous Micro-Fragmented Adipose Tissue Injection for the 81 2017 Wrong intervention 30 6 Treatment of Degenerative Knee Osteoarthritis 31 ACTRN1261900129919 82 2019 Effectiveness of Topical creams in reducing knee pain: Wrong intervention 32 0 33 Effects of treatment by Telecommunication Services on the pain, functionality and 83 RBR-662hn2 2019 Wrong intervention 34 amount of fatty tissue between muscles of patients with knee Arthrosis 35 Cyclobenzaprine HCl Extended Release 15 mg Versus Placebo in Treatment of 36 84 NCT02814565 2016 Wrong intervention Cervical and/or Lower Back Pain Due to Muscle Spasms of Local Origin 37 38 85 NCT03090152 2017 Minimal Opioid Use After Total Hip Replacement (THR) Wrong intervention 39 40 41 42 12 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 56

1 2 3 4 86 NCT02903238 2016 Brain Morphometry in OA Patients Treated With Duloxetine Wrong outcomes 5 Vestibulodynia: Understanding Pathophysiology and Determining Appropriate 87 NCT03844412 2019 Wrong population 6 Treatments 7 Three Way Interaction Between Gabapentin, Duloxetine, and Donepezil in Patients 88 NCT00619983 2008 Wrong population 8 Confidential:With Diabetic Neuropathy For Review Only 9 Pain Management in Osteoarthritis Using the Centrally Acting Analgesics Duloxetine 89 NCT02612233 2015 Wrong population 10 and Pregabalin (DUPRO) 11 A Phase 2b, Multicenter, Randomized, Double-blind, Placebo-controlled, Parallel- 12 EUCTR2019-003437- group Study to Evaluate the Efficacy, Safety, and Tolerability of MIN-101 in Patients 13 90 2014 Wrong population 42-DK With Negative Symptoms of Schizophrenia, Followed by a 24-Week, Open-Label 14 15 Extension. Doxepin Hydrochloride in Treating Oral Mucositis Pain in Patients With Head and 16 91 NCT01156142 2010 Wrong population 17 Neck Cancer Undergoing Radiation Therapy With or Without Chemotherapy Effect of of single dose amitriptiline versus pregabalin on postoperative pain reduction 18 92 IRCT201406151617N9 2015 Wrong population 19 in patients after spinal surgery 20 93 JPRN-UMIN000031428 2018 Effect of preemptive duloxetine for postoperative pain Wrong population 21 22 The Comparison of Postoperative Pain After Lumbar Fusion Surgery in Intravenous 23 94 NCT03047044 2017 Patient-controlled Analgesia Between Conventional Mode and Optimizing B.I Mode Wrong population 24 With 'PAINSTOP' Equipment Open-label Milnacipran for Persistent Knee Pain One Year After Total Knee 25 95 NCT01780389 2013 Wrong population 26 Arthroplasty (TKA) 27 96 NCT03110172 2017 Short-term Efficacy of Antidepressant in Patients Underwent TKA Wrong population 28 Effects of Duloxetine on Postoperative Wound Complication of Total Knee 29 97 NCT03880916 2019 Wrong population 30 Arthroplasty (TKA) in Central Sensitization Patients Effect of pre-operative pain treatment by means of duloxetine on postoperative 31 98 NTR4744 2014 Wrong population 32 outcome after total hip or knee arthroplasty- The DOA study 33 A randomized, multicenter, double-blind, placebo controlled, parallel study to examine EUCTR2013-004564- 34 99 2015 the efficacy and safety of a combination treatment in subjects with chronic low back Wrong intervention 55-HU 35 pain 36 An Effectiveness and Safety Study of Cyclobenzaprine HCl Alone or in Combination 100 NCT00246389 2005 Wrong intervention 37 With Ibuprofen for Acute Back or Neck Muscle Pain With Muscle Spasm 38 A Comparative Study Between Lysine Clonixinate+Cyclobenzaprine and 101 NCT01421433 2011 Wrong intervention 39 Caffeine+Carisoprodol+Sodium Diclofenac+Paracetamol 40 41 42 13 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 33 of 56 BMJ

1 2 3 Gabapentin, Methadone, and Oxycodone With or Without Venlafaxine Hydrochloride 4 102 NCT03574792 2018 in Managing Pain in Participants With Stage II-IV Squamous Cell Head and Neck Wrong intervention 5 Cancer Undergoing Chemoradiation Therapy 6 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 14 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 34 of 56

1 2 3 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

15 https://mc.manuscriptcentral.com/bmj Page 35 of 56 BMJ

1 2 3 Supplemental file 3. Calculation of effect sizes for immediate, short term and intermediate term pain in spinal pain and sciatica trials 4 Point estimate (variability), Type of Mean (SD), converted No of participants 5 Author, year Outcome Type of extracted 6 data (class) scale (range) measure Antidepressant Antidepre 7 extracted Placebo Antidepressants Placebo Placebo s ssants 8 Confidential: For Review Only 9 Pain (immediate-term) 10 Schliessbach, MD (95% NPRS (0-10) FV 0.02 (-0.51 to 0.56) 0.2 (-5.1 to 5.6) 50 50 11 2018 (TCA) CI)1 12 Konno, 2016 BPI-average Mean 13 CS -1.13 (0.11) -0.8 (0.11) -11.3 (16.6) -8 (16.5) 230 226 (SNRI) (0-10) (SE)2 14 15 Schukro, 2016 Mean VAS (0-10) 3 FV 4.35 (0.48) 6.2 (0.50) 43.5 (24) 62.1 (23.9) 25 25 16 (SNRI) (SE) Vanelderen, MD (95% 17 NPRS (0-10) FV -1.41 (-0.05 to 2.8) -11.41 (-0.5 to -27.8) 17 17 18 2015 (TCA) CI)1 19 Skljarevski, Mean NPRS (0-10) CS 1.0 (0.19) 0.34 (0.20) -10.1 (20.4) -3.4 (21.1) 111 116 20 2010b (SNRI) (SE) 21 Skljarevski, 22 2009 (SNRI) 23 BPI-average Mean 20mg CS -0.84 (0.30) -0.75 (0.22) -8.4 (22.4) -7.5 (13.5) 56 384 24 (0-10) (SE)2 25 BPI-average Mean 60mg CS -0.91 (0.22) -0.75 (0.22) -9.1 (22.8) -7.5 (13.5) 108 384 26 (0-10) (SE)2 27 BPI-average Mean 4 28 120mg 2 CS -1.22 (0.22) -0.75 (0.22) -12.2 (22.8) -7.5 (13.5) 108 37 29 (0-10) (SE) Dickens, 2000 Mean 30 VAS (0-100) FV 52.1 (25.5) 56.4 (22.3) 52.1 (25.5) 56.4 (22.3) 44 48 31 (SSRI) (SD) Hameroff, Mean 32 VAS (0-100) FV 65.2 (3.97)5 69.8 (3.73) 5 65.2 (20.6)3 69.8 (18.3) 3 27 24 33 1985 (TCA) (SE) 34 Jenkins, 1976 Mean VAS (0-10) FV 4.2 (0.65)5 4.5 (0.64) 5 42.0 (31.7)3 45 (29.3) 3 23 21 35 (TCA) (SE) 36 Pain (short-term) 37 Gould, 2020 Mean 38 DDS (0-20) FV 7.5 (5.6) 8.7 (5.5) 37.5 (28) 43.5 (27.5) 37 33 39 (TCA) (SD) 40 41 42 16 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 36 of 56

1 2 3 Urquhart, 2018 MD (95% VAS (0-100) FV -1.1 (-7.9 to 5.8) -1.1 (-7.9 to 5.8) 58 58 4 (TCA) CI) 5 Maarawi, 2018 Mean 6 VAS (0-10) FV 3.3 (1.45) 6.1 (0.92) 33.4 (14.5) 61.2 (9.2) 104 108 (TCA) (SD) 7 Schukro, 2016 Mean 8 VASConfidential: (0-10) FV 3.7 (2.90) For5.7 (2.50) Review37 (29) 57Only (25) 25 23 9 (SNRI) (SD) Konno, 2016 BPI-average Mean 10 CS -2.4 (0.11) -1.9 (0.11) -24.3 (15.9)3 -19.6 (15.5)3 209 200 11 (SNRI) (0-10) (SE) NCT01225068 Mean 12 VAS (0-100) FV 31.3 (26.4) 24.8 (32.5) 31.3 (26.4) 24.8 (32.5) 16 19 13 , 2014 (TCA) (SD) 14 Marks, 2014 VAS-Rad (0- Mean FV 13.6 (8.5) 59.7 (43.6) 13.6 (8.5) 59.7 (43.6) 7 4 15 (TCA) 100) (SD) 16 Skljarevski, BPI-average Mean CS -2.25 (0.15) -1.65 (0.15) -22.5 (20.9)3 -16.5 (21.1)3 195 199 17 2010a (SNRI) (0-10) (SE) 18 Skljarevski, BPI-average Mean CS -2.32 (0.28)7 -1.50 (0.28)7 -23.2 (21.1)3 -15 (21.1)3 111 116 19 2010b (SNRI) (0-10) (SE) 20 Skljarevski, 21 2009 (SNRI) 22 BPI-average Mean 23 20 mg CS -1.79 (0.30) -1.87 (0.22) -17.9 (22.4)3 -18.7 (13.5)3 56 384 24 (0-10) (SE) BPI-average Mean 25 60 mg CS -2.5 (0.22) -1.87 (0.22) -25.0 (22.8)3 -18.7 (13.5)3 108 384 26 (0-10) (SE) 27 BPI-average Mean 120 mg CS -2.45 (0.22) -1.87 (0.22) -24.5 (22.8)3 -18.7 (13.5)3 108 374 28 (0-10) (SE) 29 Atkinson, 2007 30 Mean 31 TCA DDS (0-20) FV 6.0 (4.1) 6.8 (3.8) 30.0 (20.5) 34.0 (19.0) 30 114 32 (SD) Mean 33 SSRI DDS (0-20) FV 7.5 (3.9) 6.8 (3.8) 37.5 (19.5) 34.0 (19.0) 31 114 34 (SD) 35 Khoromi, 2007 Mean VAS (0-10) FV 3.0 (2.7) 3.7 (2.7) 30.0 (27.0) 37.0 (27.0) 28 28 36 (TCA) (SD) 37 Katz, 2005 Mean VAS (0-10) FV 3.3 (1.9) 3.4 (1.9) 33.0 (19) 34.0 (19.0) 21 23 38 (NDRI) (SD) 39 40 41 42 17 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 37 of 56 BMJ

1 2 3 Pirbudak, 2003 Mean VAS (0-10) FV 1 (0.5) 3.3 (2.5) 10.0 (5.0) 33.0 (25.0) 30 30 4 (TCA) (SD) 5 Dickens, 2000 Mean 6 VAS (0-100) FV 57.0 (23.8) 57.0 (24.3) 57.0 (23.8) 57.0 (24.3) 44 48 (SSRI) (SD) 7 8 Atkinson, 1999 Confidential: For Review Only 9 Mean Tetracyclic DDS (0-20) FV 6.8 (4.7) 7.7 (4.6) 34.0 (23.5) 38.5 (23.1) 20 144 10 (SD) 11 Mean 12 SSRI DDS (0-20) FV 8.2 (4.0) 7.7 (4.6) 41.0 (20.0) 38.5 (23.1) 22 144 (SD) 13 Atkinson, 1998 Mean 14 DDS (0-21) CS -2.6 (4.0) -0.9 (3.4) -12.3 (19.0) -4.33 (16.3) 28 29 15 (TCA) (SD) Goodkin, 1990 Mean 16 VAS (0-10) FV 5.3 (3.0) 5.9 (2.6) 53.4 (30.0) 58.8 (26.2) 21 19 17 (SARI) (SD) Hameroff, Mean 18 VAS (0-100) FV 61.5 (5.8)5 79.6 (3.4)5 61.48 (30.2)3 79.6 (16.6)3 27 24 19 1985 (TCA) (SE) 20 Jenkins, 1976 Mean VAS (0-10) FV 3.6 (0.6) 3.8 (0.6) 36 (28.7) 38 (27.5) 23 21 21 (TCA) (SE) 22 Pain (intermediate-term) 23 Urquhart, 2018 MD (95% 24 VAS (0-100) FV -7.8 (-15.8 to 0.1) -7.8 (-15.8 to 0.1) 61 57 25 (TCA) CI) 26 Pirbudak, 2003 Mean VAS (0-10) FV 0.8 (0.5) 3.5 (2.5) 8.0 (5.0) 35.0 (25.0) 30 30 27 (TCA) (SD) 28 SD, standard deviation; MD, mean difference; 95% CI, 95% confidence interval; FV, Final Value; CS, Change Score; NPRS, Numerical Pain Rating 29 Scale; BPI, Brief Pain Inventory; VAS, Visual Analogue Scale; VAS-Rad, Visual Analogue Scale for radicular symptoms; DDS, Descriptor Differential 30 Scale; TCA, Tricyclic antidepressants, SNRI, Serotonin-Norepinephrine Reuptake inhibitors; SSRI, Selective serotonin reuptake inhibitors; NDRI, 31 Norepinephrine–dopamine reuptake inhibitor; SARI, Serotonin antagonist and reuptake inhibitors. 32 1MD (95% CI) were extracted directly when point estimate and variability measures were not available; 33 2Means extracted from figure; standard error adopted from another time point within the same study; 34 3Standard deviation estimated from standard error; 35 4 36 Sample size in the placebo group was divided by the number of groups to avoid double-counting; 5 37 Point estimate and variability measures extracted from figure; 6 38 MD (95% CI) were extracted because effect estimates reported in the paper were adjusted by baseline scores; 7 39 Standard error estimated from p-value, t-value and number of participants. 40 41 42 18 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 38 of 56

1 2 3 Supplemental file 4. Calculation of effect sizes for immediate, short and intermediate term pain in osteoarthritis trials 4 Point estimate (variability), Type of Mean (SD), converted No of participants 5 Author, year Outcome Type of extracted 6 data (class) scale (range) measure Antidepr 7 extracted Antidepressants Placebo Antidepressants Placebo Placebo essants 8 Confidential: For Review Only 9 Pain (immediate-term) 10 Uchio, 2018 BPI-average Mean CS -1.28 (0.09) -0.68 (0.09) -12.8 (11.9)1 -6.8 (11.9)1 177 176 11 (SNRI) (0-10) (SE) 12 Chappell, 2011 Mean 13 NPRS (0-10) CS -1.23 (0.20)2 -0.81 (0.18)2 -12.3 (22.6)1 -8.1 (20.3)1 1213 1273 (SNRI) (SE) 14 Frakes, 2011 BPI-average Mean 15 CS -0.6 (0.1)2 -0.3 (0.1)2 -6.0 (16.0)1 -3.0 (16.0)1 259 255 16 (SNRI) (0-10) (SE) Chappell, 2009 BPI-average Mean 17 CS -1.5 (0.16)3 0.95 (0.15)3 -15.0 (16.1)1 -10 (15.9)1 1013 1123 18 (SNRI) (0-10) (SE) 19 Pain (short-term) 20 Uchio, 2018 BPI-average Mean 21 CS -2.57 (0.12) -1.28 (0.12) -25.7 (15.9) -12.8 (15.9) 177 176 22 (SNRI) (0-10) (SE) Wang, 2017 BPI-average MD (95% 23 CS -0.50 (-0.80 to -0.20) -5.0 (-8.0 to -2.0) 172 177 24 (SNRI) (0-10) CI) Tetreault, 2016 Mean 25 VAS (0-10) FV 5.3 (1.6) 5.6 (2.0) 53.0 (16.0) 56.0 (20.0) 19 20 26 (SNRI) (SD) 27 Abou-Raya, WOMAC Mean FV 6 (4.1) 8.4 (5.4) 30.0 (20.5) 42.0 (27.0) 123 131 28 2012 (SNRI) pain (0-20) (SD) 29 Chappell, 2011 Mean NPRS (0-10) CS -2.72 (0.20)3 -1.88 (0.18)3 -27.2 (20.0)1 -18.8 (19.3)1 100 116 30 (SNRI) (SE) 31 Frakes, 2011 BPI-average Mean CS -2.46 (0.11) -1.55 (0.11) -24.6 (17.7)1 -15.5 (17.5)1 259 255 32 (SNRI) (0-10) (SE) 33 NCT01510457 MPQ-SF (0- Mean 34 FV 6.0 (5.9) 12.3 (9.6) 13.3 (13.1) 27.3 (21.3) 26 12 35 , 2012 (SNRI) 45) (SD) Chappell, 2009 BPI-average Mean 36 CS -2.93 (0.17)3 -1.98 (0.16)3 -29.3 (15.2)1 -19.8 (15.8)1 80 98 37 (SNRI) (0-10) (SE) 38 39 40 41 42 19 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 39 of 56 BMJ

1 2 3 SD, standard deviation; MD, mean difference; 95% CI, 95% confidence interval; FV, Final Value; CS, Change Score; BPI, Brief Pain Inventory; VAS, 4 Visual Analogue Scale; WOMAC, Western Ontario and McMaster Universities Osteoarthritis Index; NPRS, Numerical Pain Rating Scale; MPQ-SF, 5 McGill Pain Questionnaire – short form; SNRI, Serotonin-Norepinephrine Reuptake inhibitors. 6 1Standard deviation estimated from standard error; 7 2Means extracted from figure; standard error adopted from another time point within the same study; 8 3Data extracted from theConfidential: trial registration 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 20 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 40 of 56

1 2 3 Supplemental file 5. Calculation of effect sizes for immediate, short term and intermediate term disability in spinal pain trials 4 Point estimate (variability), Outcome Type of Mean (SD), converted No of participants 5 Type of extracted 6 Study (year) scale data measure Antidepress Antidepressant Antidepres 7 (range) extracted Placebo Placebo Placebo ants s sants 8 Confidential: For Review Only 9 Disability (immediate-term) 10 Pirbudak, ODI (0- Mean FV 36.0 (13.2) 41.0 (15.3) 36.0 (13.2) 41.0 (15.3) 30 30 11 2003 (TCA) 100) (SD) 12 13 Disability (short-term) 14 Gould, 2020 RMDQ Mean FV 7.7 (6.0) 9.5 (5.2) 31.9 (25.0) 39.6 (21.5) 37 33 15 (TCA) (0-24) (SD) 16 Urquhart, RMDQ MD (95% FV -1.6 (-2.9 to -0.4)1 -7.0 (-12.5 to -1.5) 58 58 17 2018 (TCA) (0-23) CI) 18 Maarrawi, NDI (0- Mean 19 CS -42.2 (15.5) -13.7 (9.6) -42.2 (15.5) -13.7 (9.6) 104 108 2018 (TCA) 100) (SD) 20 Konno, 2016 RMDQ Mean 21 CS -3.9 (0.22) -3.2 (0.22) -16.1 (13.2)2 -13.5 (13.0) 2 230 226 22 (SNRI) (0-24) (SE) Marks, 2014 ODI (0- Mean 23 FV 18.1 (8..0) 22.5 (15.0) 18.1 (8..0) 22.5 (15.0) 5 3 24 (SNRI) 100) (SD) Skljarevski 25 RMDQ Mean 26 (2010)a CS -2.7 (0.31) -2.2 (0.32) -11.2 (17.2)2 -9.2 (17.8) 2 178 179 (0-24) (SE) 27 (SNRI) 28 Skljarevski RMDQ Mean 29 (2010)b CS -3.6 (4.3)3 -1.9 (4.3)3 -15.0 (17.8)2 -8.0 (17.8) 2 109 116 (0-24) (SE) 30 (SNRI) 31 Skljarevski, 32 2009 (SNRI) 33 RMDQ Mean 34 20mg CS -2.3 (4.3)4 -1.3 (4.3)4 -9.4 (17.8) -5.4 (17.8) 56 38 (0-24) (SD) 35 RMDQ Mean 36 60mg CS -2.7 (4.3)4 -1.3 (4.3)4 -11.2 (17.8) -5.4 (17.8) 108 38 37 (0-24) (SD) RMDQ Mean 38 120mg CS -2.8 (4.3) -1.3 (4.3)4 -11.7 (17.8) -5.4 (17.8) 108 37 39 (0-24) (SD) 40 41 42 21 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 41 of 56 BMJ

1 2 3 Atkinson, RMDQ Mean FV 2.3 (2.5)5 4.1 (2.5)5 9.6 (10.4) 17.1 (10.4) 30 11 4 2007 (TCA) (0-24) (SD) 5 Khoromi, ODI (0- Mean 6 FV 27.5 (16.7) 30.5 (15.9) 27.5 (16.7) 30.5 (15.9) 28 28 2007 (TCA) 100) (SD) 7 Pirbudak, ODI (0- Mean 8 Confidential:FV 22.2 (11.2) For35.0 (14.0) Review22.2 (11.2) 35.0 Only (14.0) 30 30 9 2003 (TCA) 100) (SD) Dickens, ODI (0- Mean 10 FV 50.2 (15.2) 44.0 (52.4) 50.2 (15.2) 44.0 (52.4) 44 48 11 2000 (SSRI) 100) (SD) Goodkin, SIP (0- Mean 12 FV 25.4 (17.0) 22.8 (12.7) 25.4 (17.0) 22.8 (12.7) 21 19 13 1990 (SARI) 100) (SD) 14 Disability (intermediate-term) 15 Urquhart, RMDQ MD (95% 16 FV -0.98 (-2.4 to 0.5)1 -4.3 (-10.5 to 2.0) 61 57 17 2018 (TCA) (0-23) CI) Pirbudak, ODI (0- MD (95% 18 FV 18.0 (12.0) 38.0 (18.0) 18.0 (12.0) 38.0 (18.0) 30 30 19 2003 (TCA) 100) CI) 20 SD, standard deviation; 95% CI, 95% confidence interval; FV, Final Value; CS, Change Score; ODI, Oswestry Disability Index; RMDQ, Roland- 21 Morris Disability Questionnaire; SIP, Sickness Impact Profile; TCA, Tricyclic antidepressant; SNRI, Serotonin-norepinephrine receptor inhibitor; 22 SSRI, Selective serotonin reuptake inhibitors; SARI, Serotonin receptor antagonists and reuptake inhibitors. 23 1MD (95% CI) were extracted because effect estimates reported in the paper were adjusted by baseline scores); 24 2Standard deviation estimated from standard error; 25 3Standard error estimated from p-value, t-value and number of participants; 26 4Standard deviation from Skljarevski (2010)b 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 22 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 42 of 56

1 2 3 Supplemental file 6. Calculation of effect sizes for immediate, short term and intermediate term disability in osteoarthritis trials 4 Point estimate (variability), Outcome Type of Mean (SD), converted No of participants 5 Type of extracted 6 Study (year) scale data measure Antidepressant Antidepressant Antidepr 7 (range) extracted Placebo Placebo Placebo s s essants 8 Confidential: For Review Only 9 Disability (immediate-term) 10 WOMAC Mean 11 Uchio (2018) -Total (0- CS -10.1 (0.75) -5.08 (0.74) -10.4 (10.6)1 -5.3 (10.6)1 177 176 (SE) 12 96) 13 14 Disability (short-term) 15 WOMAC Mean -10.9 16 Uchio (2018) -Total (0- CS -17.41 (0.91) -10.45 (0.91) -18.1 (12.0)1 177 176 (SE) (12.0)1 17 96) 18 WOMAC 19 MD (95% Wang (2017) -Total (0- CS -3.5 (-5.6 to -1.4)2 -3.6 (-5.8 to -1.5) 184 182 20 CI) 21 96) WOMAC 22 Tetreault Mean -Total (0- FV 39.9 (17.4) 39.1 (18.9) 41.5 (18.1) 40.7 (19.7) 19 20 23 (2016) (SD) 24 96) 25 WOMAC Abou-Raya -Physical Mean 26 FV 24.6 (8.4) 30.3 (9.8) 36.2 (12.3) 44.5 (14.4) 123 131 27 (2012) function (SD) 28 (0-68) 29 WOMAC Chappell Mean -14.3 (17.3) 30 -Total (0- CS -17.5 (1.5) -13.7 (1.5) -18.2 (18.1)1 128 128 (2011) (SE) 1 31 96) 32 WOMAC 33 Frakes -Physical Mean -14.4 CS -21.3 (1.1) -14.4 (1.1) -21.3 (17.3) 1 251 253 34 (2011) function (SE) (17.7)1 35 (0-100) 36 NCT015104 PDI (0- Mean 37 FV 15.0 (10.0) 26.0 (18.0) 21.4 (14.3) 37.1 (25.7) 26 12 38 57 (2012) 70) (SD) 39 SD, standard deviation; 95% CI, 95% confidence interval; FV, Final Value; CS, Change Score; WOMAC, Western Ontario and McMaster 40 Universities Osteoarthritis Index; PDI, Pain Disability Index; SNRI, Serotonin-norepinephrine receptor inhibitor; 41 42 23 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 56 BMJ

1 2 3 1Standard deviation estimated from standard error; 4 2MD (95% CI) were extracted directly when point estimate and variability measures were not readily available 5 6 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 24 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 56

1 2 3 4 Supplemental file 7. Number of serious adverse events, adverse events, and drop-outs due to adverse events and in spinal pain 5 and osteoarthritis trials. 6 Serious adverse events, n/N Any adverse event, n/N Drop-outs due to adverse events, n/N 7 Study (year) Antidepressants Placebo Antidepressants Placebo Antidepressants Placebo 8 Confidential: For Review Only 9 Spinal pain 10 Gould, 2020 1/381 0/331 12/171 6/171 3/381 2/331 11 (TCA) 12 Schliessbach, 13 NR NR NR NR NR NR 2018 (TCA) 14 Urquhart, 2018 15 NR NR 25/72 30/74 9/72 7/74 16 (TCA) Maarrawi, 2018 17 NR NR 8/112 0/108 8/112 0/108 18 (TCA) 19 Schukro, 2016 NR NR 20/31 18/29 NR NR 20 (SNRI) 21 Konno, 2016 4/232 4/226 167/2341 134/2241 16/232 8/226 22 (SNRI) 23 Vanelderen, 2015 0/20 0/20 2/20 0/20 2/20 0/20 24 (TCA) 25 NCT01225068, 0/20 1/20 14/201 10/201 NR NR 26 2014 (SNRI) 27 Marks, 2014 28 0/81 0/51 0/81 0/51 2/81 0/51 29 (SNRI) Skljarevski, 2010a 30 5/1981 0/2031 123/1981 112/2031 30/1981 11/2031 31 (SNRI) Skljarevski 32 5/1151 1/1211 72/1151 59/1211 16/1151 7/1211 33 (2010)b (SNRI) 34 Skljarevski, 2009 35 (SNRI) 36 20 mg 1/591 1/381,2 39/591 23/381,2 9/591 3.3/381,2 37 38 60 mg 1/1161 1/381,2 78/1161 23/381,2 17/1161 3.3/381,2 39 40 41 42 25 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 56 BMJ

1 2

3 1 1,2 1 1,2 1 1,2 4 120 mg 3/112 1/37 80/112 23/38 27/112 3.3/37 5 Atkinson, 2007 6 7 TCA NR NR 19/30 1.5/11 17/52 0.5/13 8 Confidential: For Review Only 9 SSRI NR NR 16/31 1.5/11 3/43 0.5/13 10 Khoromi, 2007 NR NR 19/28 14/28 2/55 1/55 11 (TCA) 12 Katz, 2005 2/44 1/44 NR3 NR3 3/54 3/54 13 (NDRI) 14 Pirbudak, 2003 0/30 0/30 0/30 0/30 0/30 0/30 15 (TCA) 16 Dickens, 2000 17 NR NR NR NR NR NR 18 (SSRI) 19 Atkinson, 1999 20 2 21 TCA NR NR 18/20 15.5/16 10/33 0.5/18 22 SSRI NR NR 20/22 15.5/16 6/22 0.5/182 23 Atkinson, 1998 24 NR NR 38/38 40/40 5/28 4/29 25 (TCA) Goodkin, 1990 26 3/22 1/20 NR NR 3/22 1/20 27 (SARI) 28 Hameroff, 1985 NR NR NR NR 3/30 3/30 29 (TCA) 30 Pheasant, 1983 NR NR 1/9 2/9 0/16 1/16 31 (TCA) 32 Alcoff, 1982 NR NR 22/28 NR 5/28 0/22 33 (TCA) 34 Hameroff, 1982 35 NR NR NR NR 1/15 2/15 36 (TCA) Jenkins, 1976 37 NR NR NR NR 1/23 1/21 38 (TCA) 39 Osteoarthritis 40 41 42 26 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 56

1 2 3 Uchio, 2018 1/1781 1/1761 120/1781 98/1761 11/1781 2/1761 4 (SNRI) 5 Wang, 2017 6 0/1991 3/1981 122/1991 84/1981 20/2051 10/2021 (SNRI) 7 Tetreault, 2016 8 Confidential:0/191 0/211 6/19 1For Review6/211 NR OnlyNR 9 (SNRI) Abou-Raya, 2012 10 0/144 0/144 NR NR 9/144 6/144 11 (SNRI) Chappell, 2011 12 1/1111 2/1201 55/1111 49/1201 9/1111 5/1201 13 (SNRI) 14 Frakes, 2011 5/2641 3/2601 167/2641 129/2601 40/2641 23/2601 15 (SNRI) 16 NCT01510457, 0/26 0/12 9/26 0/12 NR NR 17 2012 (SNRI) 18 Chappell, 2009 3/128 2/128 64/128 41/128 24/128 7/128 19 (SNRI) 20 21 TOTAL 35/1,836 22/1,662 1,276/2,171 895/1,936 311/2,533 116/2,257 22 SNRI, Serotonin-norepinephrine receptor inhibitor; SSRI, Selective serotonin reuptake inhibitor; TCA, Tricyclic antidepressant; 23 NDRI, Norepinephrine and dopamine reuptake inhibitor 24 1Data extracted from the trial registration; 25 2Sample size in the placebo group was divided by the number of groups to avoid double-counting; 26 3Data reported on the number of adverse events experiences by at least 2 participants 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 27 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 56 BMJ

1 2 3 Supplemental file 8. GRADE Framework 4 5 6  Risk of bias: downgrade by one level if more than 25% of participants are from studies at high 7 8 risk of bias1 (i.e. one or more bias domains were judged high-risk) 9 2 10  Inconsistency: downgrade by one level if heterogeneity was large (I statistic value >50%, 11 representing potentially substantial heterogeneity)2 12 Confidential: For Review Only 13  Indirectness: We will not assess indirectness as patients, interventions and comparators will be 14 similar across studies. 15 16  Imprecision: 17 18 o Continuous outcomes: downgrade by one level if the limits of the 95% confidence 19 interval are 20 points different to the point estimate on a 100-point scale. 20 21 o Dichotomous outcomes. downgrade by one level if the lower or upper limits of the 95% 22 confidence interval include appreciable benefit or harm (i.e 95% CI under 0.75 or over 23 24 1.25).4 25 26  Small study effects: downgrade by one level if a funnel plot suggests the presence of publication 27 bias,5 or if more than 25% of participants were from small studies (< 100 participants per arm). 28 29 30 REFERENCES 31 1. Guyatt GH, Oxman AD, Vist G, et al. GRADE guidelines: 4. Rating the quality of evidence-- 32 study limitations (risk of bias). J Clin Epidemiol. 2011;64(4):407-415. 33 2. Guyatt GH, Oxman AD, Kunz R, et al. GRADE guidelines: 7. Rating the quality of evidence- 34 -inconsistency. J Clin Epidemiol. 2011;64(12):1294-1302. 35 36 3. Guyatt GH, Oxman AD, Kunz R, et al. GRADE guidelines: 8. Rating the quality of evidence- 37 -indirectness. J Clin Epidemiol. 2011;64(12):1303-1310. 38 4. Guyatt GH, Oxman AD, Kunz R, et al. GRADE guidelines 6. Rating the quality of evidence-- 39 imprecision. J Clin Epidemiol. 2011;64(12):1283-1293. 40 5. Guyatt GH, Oxman AD, Montori V, et al. GRADE guidelines: 5. Rating the quality of 41 evidence-publication bias. J Clin Epidemiol. 2011;64(12):1277-1282 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

28 https://mc.manuscriptcentral.com/bmj BMJ Page 48 of 56

1 2 3 Supplemental file 9. Characteristics of the included studies (n = 33) 4 Industry Outcomes and time 5 Author, year Characteristics of participants Interventions and dose regimens 6 funding points 7 Spinal pain (25 trials, 29 comparisons) 8 Confidential: ForDrug (class): Review Desipramine hydrochloride (TCA) Only Pain (DDS, 0-20), 9 Condition: Chronic low back pain (≥ 6 months) Dose: dose targeted to achieve a serum concentration of 5- disability (RMDQ, 0- 10 Mean age (SD): intervention 55.6 (11.7), control 57.9 (10.9) 60 ng/ml (20-60 mg/day) 24) at 12 weeks. 11 Gould 202051 N (%) female: intervention 5 (13.5), control 3 (9.7) Duration: 12 weeks No Adverse events were 12 Neuropathic pain: not mentioned Control: Active placebo (Benztropine) 0.125 mg/day, single monitored throughout 13 Depression: No dose, for 12 weeks study period. 14 Sample size (IG/PG): 37/33 15 Drug (class): Imipramine (TCA) 16 Dose: 75 mg/day; Condition: Chronic low back pain (≥ 3 months); Duration: single treatment; 17 Schliessbach, Mean age (SD): 54.4 (17.3); Sample size (IG/PG): 25/25 Pain (NPRS, 0-10) at 2 18 2018 (cross N (%) female: 32 (64); No Control: Active placebo (Tolterodine) 1mg/day, single hours post-intervention 19 over trial)28 Neuropathic pain: Not mentioned; dose; single treatment; 20 Depression: No 21 Rescue medication: allowed 22 Washout period: 1 week; Drug (class): Low-dose Amitriptyline (TCA) 23 Pain (VAS, 0-100), Condition: Chronic low back pain (≥ 3 months) Dose: 25 mg/day 24 disability (RMDQ, 0- Mean age (SD): Intervention 53.5 (14.2), control 56.0 (13.2); Duration: 26 weeks 25 Urquhart, 23) at 12 and 26 weeks. N (%) female: Intervention 28 (39), control 28 (38); Sample size: 72/74 No 26 201852 Adverse events were Neuropathic pain: Not mentioned Control: Active placebo (Benztropine) 1 mg/day, single 27 monitored throughout Depression: No dose, for 26 weeks 28 study period Rescue medication: Not mentioned 29 Drug (class): Low-dose Amitriptyline (TCA) Pain (VAS, 0-10), 30 Condition: Chronic neck pain (≥ 3 months); Dose: 5 mg/day disability (NDI, 0-100) Mean age (SD): intervention 43.5 (11.1), control 45.1 (11.5); 31 Maarrawi, Duration: 8 weeks at 8 weeks. Adverse N (%) female: Intervention 97 (86.6), control 93 (86.1) No 32 201846 Sample size (IG/PG): 112/108 events were monitored Neuropathic pain: Not mentioned 33 Control: Inert placebo during the first 12 days Depression: No 34 Rescue medication: Not mentioned of treatment 35 Condition: Chronic sciatica (> 6 months); Drug (class): Duloxetine (SNRI) Pain (VAS, 0-10) at 2 36 Schukro, 2016 Mean age (SD): 57.9 (13.4); Dose: 120 mg/day and 4 weeks; adverse 37 (cross over N (%) female: 21 (51.0); Duration: 4 weeks No events were monitored 38 trial)29 Neuropathic pain: Neuropathic pain: Yes (clinical features + Sample size (IG/PG): 31/29 throughout study 39 painDETECT ≥ 12) Control: Inert placebo period 40 41 42 29 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 56 BMJ

1 2 3 Depression: No Rescue medication: Allowed 4 Washout period: 2 weeks 5 Drug (class): Duloxetine (SNRI) Pain (BPI, 0-10), Condition: Chronic low back pain (≥ 6 months); 6 Dose: 60 mg/day disability (RMDQ, 0- Mean age (SD): Intervention 57.8 (13.7), control 60.0 (13.2); 7 Duration: 14 weeks 24) at 14 weeks. Konno, 201633 N (%) female: Intervention 122 (54.0), control 115 (50.0); Yes 8 Sample size (IG/PG): 232/226 Adverse events were NeuropathicConfidential: pain: Not mentioned For Review Only 9 Control: Inert placebo monitored throughout Depression: No 10 Rescue medication: Not mentioned study period 11 Drug (class): Amitriptyline (TCA) Condition: Sciatica (duration not reported); Pain (NPRS, 0-10) at 2 Dose: 25 mg/day 12 Mean age (SD): Intervention 50.0 (3.0), control 51.0 (3.0); weeks. Adverse events Vanelderen, Duration: 2 weeks 13 N (%) female: Intervention 7 (35), control 11 (55); No were monitored 201553 Sample size (IG/PG): 20/20 14 Neuropathic pain: Yes (clinical features + DN4 ≥ 4); throughout study Control: Inert placebo 15 Depression: No period. 16 Rescue medication: Allowed Drug (class): Milnacipran (SNRI) 17 Condition: Chronic sciatica (≥ 6 months); Pain (VAS, 0-100) at 6 Dose: 100 mg/day 18 Mean age (SD): intervention 47.1 (11.6), control 48.3 (9.1); weeks. Adverse events NCT01225068 Duration: 6 weeks 19 N (%) female: Intervention 10 (50), control 11 (55); No were monitored , 201454 Sample size (IG/PG): 20/20 20 Neuropathic pain: Yes (clinical features) throughout study Control: Inert placebo 21 Depression: No period (8 weeks). Rescue medication: Not mentioned 22 Pain (VAS, 0-100), 23 Drug (class): Milnacipran (SNRI) Condition: Chronic sciatica (≥ 6 months); disability (ODI, 0-100) 24 Dose: 100 mg/day Mean age (SD): Not reported at 10 weeks. Adverse 25 Duration: 10 weeks Marks, 201434 N (%) female: Not reported Yes events were monitored Sample size (IG/PG): 7/4 26 Neuropathic pain: Yes (clinical features) (measurement Control: Inert placebo 27 Depression: No timeframe not Rescue medication: Not allowed 28 specified) 29 Drug (class): Duloxetine (SNRI) Pain (BPI, 0-10), Condition: Chronic low back pain (≥ 6 months); 30 Dose: 60 mg/day disability (RMDQ, 0- Mean age (SD): Intervention 54.9 (13.7), control 53.4 (14.2); 31 Skljarevski, Duration: 12 weeks 24) at 12 weeks. N (%) female: Intervention 118 (59.6), control 128 (63.1) Yes 32 2010a35 Sample size (IG/PG): 198/203 Adverse events were Neuropathic pain: No 33 Control: Inert placebo monitored throughout Depression: No 34 Rescue medication: Allowed study period. 35 Condition: Chronic low back pain (≥ 6 months); Drug (class): Duloxetine (SNRI) Pain (BPI, 0-10), 36 Mean age (SD): Mean age (SD): intervention 51.8 (14.9), Dose: 60-120 mg/day Skljarevski, disability (RMDQ, 0- 37 control 51.2 (13.5); Duration: 13 weeks Yes 2010b37 24) at 13 weeks. 38 N (%) female: Intervention 71 (61.7), control 73 (60.3) Sample size (IG/PG): 115/121 Adverse events were 39 Neuropathic pain: No Control: Inert placebo 40 41 42 30 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 56

1 2 3 Depression: No Rescue medication: Allowed monitored throughout 4 study period. 5 Condition: Chronic low back pain (≥ 6 months); 6 Mean age (SD): Intervention (20 mg/day) 52.9 (12.8), Drug (class): Duloxetine (SNRI) Pain (BPI, 0-10), 7 intervention (60 mg/day) 53.3 (14.7), intervention (60 Dose: 20, 60 and 120 mg/day disability (RMDQ, 0- 8 mg/day) 54.9 (14.8), control 54.0 (13.5); Duration: 13 weeks Skljarevski, Confidential: For Review Only 24) at 12 weeks. 9 N (%) female: Intervention (20 mg/day) 36 (61.0), Sample size (IG/PG): 59 (20mg/day)/116 (60mg/day), 112 Yes 200936 Adverse events were 10 intervention (60 mg/day) 67 (57.8), intervention (60 mg/day) (120 mg/day)/117 monitored throughout 65 (58.0), control 64 (54.7); Control: Inert placebo 11 study period 12 Neuropathic pain: No Rescue medication: Allowed; 13 Depression: No 14 Drug (class): Desipramine hydrochloride (TCA) 15 Dose: Target dose between 50 and 150 ng/ml 16 Duration: 12 weeks 17 Pain (DDS, 0-20), Condition: Chronic low back pain (≥ 6 months); Drug (class): Fluoxetine (SSRI) 18 disability (RMDQ, 0- Mean age (SD): 46.4 (10.2); Dose: Target dose between 100 to 400 ng/ml 19 Atkinson, 24) at 12 weeks. N (%) female: 47 (38.8); Duration: 12 weeks No 20 200755 Adverse events were Neuropathic pain: Not mentioned 21 monitored throughout Depression: No Sample size: (TCA/SSRI/PG): 30/31/22 22 study period. 23 24 Control: Active placebo (Benztropine 0.5 mg/day) 25 26 Rescue medication: Not mentioned Drug (class): Nortriptyline (Tricyclic) 27 Pain (VAS, 0-10), Condition: Chronic sciatica (≥ 3 months); Dose: 100 mg/day 28 disability (ODI, 0-100) Khoromi, 2007 Median age (range): 53 (19-65); Duration: 4 weeks 29 at 5 weeks. Adverse (cross over N (%) female: 25 (83.3) Sample size (IG/PG): 28/28 No 30 events were monitored trial)30 Neuropathic pain: Yes (clinical features) Control: Inert placebo 31 throughout study Depression: No Rescue medication: Not mentioned 32 period. Washout period: 2 weeks 33 Drug (class): Bupropion (Norepinephrine-dopamine- 34 Pain (VAS, 0-10) at 6 Condition: Chronic low back pain (≥ 3 months) 35 reuptake inhibitor) weeks. Adverse events Katz, 2005 Mean age (SD): Intervention 49.8 (10.0), control 51.4 (11.4); 36 Dose: 300 mg/day were monitored (cross over N (%) female: Intervention 12 (57.1), control 9 (39.1); Yes 37 Duration: 7 weeks (measurement trial)31 Neuropathic pain: No Sample size (IG/PG): 52/47 38 timeframe not Depression: Not mentioned Control: Inert placebo 39 specified). Washout period: 1 week 40 41 42 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 51 of 56 BMJ

1 2 3 Drug (class): Amitriptyline (Tricyclic) (in addition to Pain (VAS, 0-10), 4 epidural corticosteroid injection) disability (ODI, 0-100) Condition: Acute sciatica (≤ months); 5 Dose: 50 mg/day at 2, 12 and 26 weeks. Mean age (SD): Intervention 35.1 (4.4), control 35.7 (7.5); 6 Pirbudak, Duration: 26 weeks Adverse events were N (%) female: Intervention 19 (63.3), control 24 (80.0) No 7 200345 Sample size (IG/PG): 30/30 monitored Neuropathic pain: Yes (clinical features) 8 Control: Inert placebo (in addition to epidural corticosteroid (measurement DepressionConfidential:: Not mentioned For Review Only 9 injection timeframe not 10 Rescue medication: Not mentioned specified). Drug (class): Paroxetine (SSRI) 11 Condition: Chronic low back pain (≥ 6 months); Pain (VAS, 0-100), Dose: 20 mg/day 12 Mean age (SD): Intervention 44 (9.7), control 46.0 (10.6); disability (ODI, 0-100) Dickens, Duration: 8 weeks 13 N (%) female: Intervention 23 (52.3), control 27 (56.3); Yes at 2 and 8 weeks. 200038 Sample size (IG/PG): 44/48 14 Neuropathic pain: No Adverse events were Control: Inert placebo 15 Depression: Yes (MADRS ≥ 16) not reported. 16 Rescue medication: Allowed 17 Drug (class): Maprotiline (Tetracyclic) 18 Dose: 150 mg/day Duration: 8 weeks Pain (DDS, 0-20) at 8 19 Condition: Chronic low back pain (≥ 6 months); Drug (class): Paroxetine (SSRI) weeks. Adverse events 20 Mean age (SD): 49.2 (9.4); Atkinson, Dose: 30 mg/day were reported 21 N (%) female: 38 (36.9); No 199956 Duration: 8 weeks (measurement 22 Neuropathic pain: No timeframe not 23 Depression: No Sample size (Tetracyclic/SSRI/PG): 33/34/36 specified). 24 Control: Active placebo (Diphenhydramine 37.5 mg/day) 25 Rescue medication: Not mentioned 26 Drug (class): Nortriptyline (TCA) Pain (DDS, 0-21) at 8 27 Condition: Chronic low back pain (≥ 6 months); Dose: 100 mg/day weeks. Adverse events 28 Mean age (SD): Intervention 45.7 (10.6), control 47.1 (10.6); Duration: 8 weeks Atkinson, were monitored 29 N (%) female: Not reported; Sample size (IG/PG): 38/40 No 199857 (measurement 30 Neuropathic pain: Not mentioned; Control: Inert placebo timeframe not 31 Depression: No specified). 32 Rescue medication: Not mentioned 33 Condition: Chronic low back pain (≥ 12 months or acute low Pain (VAS, 0-10), Drug (class): Trazodone (SARI) 34 back pain ≥ 2 weeks duration) disability (SIP, 0-100) Dose: 600 mg/day 35 Goodkin, Mean age (SD): Intervention 51.4 (13.1), control 56.1 (12.6); at 6 weeks. Only Duration: 6 weeks Yes 36 199039 N (%) female: Intervention 8 (36.4), control 8 (40.0); hospitalisation during Sample size (IG/PG): 22/20 37 Neuropathic pain: Not mentioned; the trial was reported Control: Inert placebo 38 Depression: 40% had clinical depression comorbid to pain as adverse event. 39 40 41 42 32 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 56

1 2 3 Condition: Chronic low back and/or neck pain (duration not Drug (class): Doxepin (TCA) Pain (VAS, 0-100) at 6 4 reported); Dose: 300 mg/day weeks. Adverse events 5 Hameroff, Mean age (SD): Intervention 48.9 (2.4), control 48.4 (2.0); Duration: 6 weeks Unclear were not monitored 6 198558 N (%) female: Intervention 13 (43.3), control 15 (50.0); Sample size (IG/PG): 30/30 throughout study 7 Neuropathic pain: Not mentioned Control: Inert placebo period. 8 DepressionConfidential:: Yes (HDS ≥ 18) ForRescue medication Review: Not mentioned Only 9 Drug (class): Amitriptyline (TCA) Condition: Chronic low back pain (≥ 12 months); Disability (Functional 10 Dose: 150 mg/day Pheasant 1983, Mean age (SD): 47.2 (22-68); evaluation score, 0-4). Duration: 6 weeks 11 cross over N (%) female: 4 (25.0); Unclear Adverse events Sample size (IG/PG): 16/16 12 trial32 Neuropathic pain: Not mentioned monitoring was Control: Active placebo (Atropine 0.2 mg/day); 13 Depression: Not mentioned unclear. 14 Washout period: 2 weeks 15 Condition: Chronic low back pain (duration not reported); Drug (class): Imipramine (TCA) Mean age (SD): Intervention 29.2, control 33.8; Dose: 150 mg/day 16 Adverse events were N (%) female: Intervention 14 (50.0), control 12 (54.5); Duration: 8 weeks 17 Alcoff, 198259 Unclear monitored throughout Neuropathic pain: Not mentioned Sample size (IG/PG): 28/22 18 study period. 19 Depression: Yes (clinical diagnosis) in 6 (21.4) and 4 (18.2) Control: Inert placebo 20 participants in group intervention and control, respectively. Rescue medication: Not mentioned 21 Condition: Chronic neck and/or low back pain (≥ 2 Drug (class): Doxepin (TCA) months); Dose: 50-300 mg/day Pain (VAS, 0-100) at 2 22 Hameroff, Mean age (SD): 46.6 (2.3); Duration: 6 weeks and 6 weeks. Unclear 23 Unclear 198260 N (%) female: 15 (50); Sample size (IG/PG): 15/15 how adverse events 24 Neuropathic pain: Not mentioned Control: Inert placebo were monitored. 25 Depression: Yes (clinical depression) Rescue medication: Not mentioned 26 Condition: Chronic low back pain (duration not reported); Drug (class): Imipramine (TCA) 27 Mean age (SD): Intervention 26.0 (18-45), control 26.7 (18- Pain (VAS, 0-10) at 2 Dose: 75 mg/day 28 49); and 4 weeks. Adverse Duration: 4 weeks 29 Jenkins, 197640 N (%) female: Intervention 0 (0), control 3 (14.3). Yes events were monitored Sample size (IG/PG): 23/21 30 Neuropathic pain: Not mentioned at the end of the study Control: Inert placebo 31 Depression: Half of participants in each group had a clinical period. Rescue medication: Not mentioned 32 diagnosis of depression. 33 Osteoarthritis (8 trials, 8 comparisons) 34 Drug (class): Duloxetine (SNRI) Pain (BPI, 0-11), Condition: Knee osteoarthritis (ACR criteria ≥ 3 months); 35 Dose: 60 mg/day disability (WOMAC Mean age (SD): Intervention 65.5 (8.0), control 66.4 (8.4); 36 Duration: 14 weeks total score, 0-96). Uchio, 201841 N (%) female: Intervention 142 (80.2), control 132 (75.0); Yes 37 Sample size (IG/PG): 177/177 Adverse events were Neuropathic pain: Not mentioned 38 Control: Inert placebo monitored Depression: Not mentioned 39 Rescue medication: Allowed (measurement 40 41 42 33 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 56 BMJ

1 2 3 timeframe not 4 specified). 5 Pain (BPI, 0-11), 6 Drug (class): Duloxetine (SNRI) disability (WOMAC Condition: Knee osteoarthritis (ACR criteria ≥ 3 months); 7 Dose: 60 mg/day total score, 0-96). Mean age (SD): Intervention 65.5 (8.0), control 66.4 (8.4); 8 Duration: 14 weeks Adverse events were Uchio, 201841 N (%) femaleConfidential:: Intervention 142 (80.2), control 132 (75.0); For Review OnlyYes 9 Sample size (IG/PG): 177/177 monitored Neuropathic pain: Not mentioned 10 Control: Inert placebo (measurement Depression: Not mentioned 11 Rescue medication: Allowed timeframe not 12 specified). 13 Pain (BPI, 0-10), Drug (class): Duloxetine (SNRI) disability (WOMAC 14 Condition: Knee osteoarthritis (ACR criteria ≥ 3 months); Dose: 60 mg/day total score, 0-96) at 13 15 Mean age (SD): Intervention 61.2 (8.2), control 59.8 (8.4); Duration: 13 weeks weeks. Adverse events 16 Wang, 201724 N (%) female: Intervention 160 (78.0), control 151 (74.8); Yes Sample size (IG/PG): 205/202 were monitored 17 Neuropathic pain: Not mentioned Control: Inert placebo (measurement 18 Depression: No 19 Rescue medication: Allowed timeframe not 20 specified). 21 Pain (VAS, 0-10), Condition: Knee osteoarthritis (ACR criteria ≥ 12 months); Drug (class): Duloxetine (SNRI) disability (WOMAC 22 Mean age (SD): Intervention 57.6 (0.8) to 60.3 (1.7), control Dose: 60 mg/day total score, 0-96) at 12 23 Tetreault, 54.7 (3.0) to 62.0 (2.6); Duration: 12 weeks weeks. Adverse events 24 No 201661 N (%) female: Intervention 10 (52.6), control 12 (57.1); Sample size (IG/PG): 19/20 were monitored 25 Neuropathic pain: Not mentioned Control: Inert placebo (measurement 26 Depression: No Rescue medication: Allowed; timeframe not 27 specified). 28 Pain (WOMAC pain 29 subscale, 0-20), Drug (class): Duloxetine (SNRI) 30 Condition: Knee osteoarthritis (ACR criteria ≥ 3 months); disability (WOMAC Dose: 60 mg/day 31 Mean age (SD): Intervention 68.9 (6.2), control 68.5 (5.8); physical function Abou-Raya, Duration: 14 weeks 32 N (%) female: Intervention 121 (84.0), control (83.0); No subscale, 0-68) at 16 201262 Sample size (IG/PG): 144/144 33 Neuropathic pain: Not mentioned weeks. Adverse events Control: Inert placebo 34 Depression: Yes (GDS ≥ 5) were monitored Rescue medication: Allowed 35 throughout study 36 period. 37 Condition: Knee osteoarthritis (ACR criteria ≥ 3 months); Drug (class): Duloxetine (SNRI) Pain (BPI, 0-10), Chappell, Mean age (SD): Intervention 63.2 (8.8), control 61.9 (9.2); Dose: 60 mg/day Yes disability (WOMAC 38 201143 39 N (%) female: Intervention 89 (69.5), control 107 (83.6); Duration: 13 weeks total score, 0-96) at 13 40 41 42 34 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 56

1 2 3 Neuropathic pain: Not mentioned Sample size (IG/PG): 128/128 weeks. Adverse events 4 Depression: Not mentioned Control: Inert placebo were monitored 5 Rescue medication: Allowed throughout study 6 period. 7 Pain (BPI, 0-10), 8 Drug (class): Duloxetine (SNRI) disability (WOMAC ConditionConfidential:: Knee osteoarthritis (ACR criteria ≥ 3 months); For Review Only 9 Dose: 120 mg/day physical function Mean age (SD): Intervention 61.6 (9.2), control 60.3 (9.2); 10 Duration: 10 weeks subscale, 0-100) at 8 Frakes, 201144 N (%) female: Intervention 152 (57.6), control 147 (56.5); Yes Sample size (IG/PG): 264/260 weeks. Adverse events 11 Neuropathic pain: Not mentioned Control: Inert placebo were monitored 12 Depression: Not mentioned 13 Rescue medication: Allowed throughout study 14 period. Pain (MPQ-SF, 0-45), 15 Drug (class): Milnacipran (SNRI) Condition: Knee osteoarthritis (ACR criteria ≥ 6 months); disability (PDI, 0-70) 16 Dose: 200 mg/day Mean age (SD): Intervention 57.0 (9.0), control 53.0 (7.0); at 11 weeks. Adverse 17 NCT01510457 Duration: Approximately 8 weeks N (%) female: Intervention 17 (65.4), control 6.0 (50.0); Unclear events were monitored 18 201263 Sample size (IG/PG): 29/17 Neuropathic pain: Not mentioned (measurement 19 Control: Inert placebo Depression: Not mentioned timeframe not 20 Rescue medication: Not mentioned specified). 21 Pain (BPI, 0-10), 22 Drug (class): Duloxetine (SNRI) disability (WOMAC 23 Condition: Knee osteoarthritis (ACR criteria ≥ 3 months); Dose: 120 mg/day physical function 24 Mean age (SD): Intervention 62.1 (9.6), control 62.5 (9.3); Chappell Duration: 13 weeks subscale, 0-68) at 12 N (%) female: Intervention 70 (63.1), control 81 (67.5); Yes 25 200942 Sample size (IG/PG): 111/120 weeks. Adverse events Neuropathic pain: Not mentioned 26 Control: Inert placebo were monitored Depression: No 27 Rescue medication: Allowed throughout study 28 period. 29 Legend: NPRS, Numerical Pain Rating Scale; VAS, Visual Analogue Scale; RMDQ, Roland-Morris Disability Questionnaire; NDI, Neck Disability Index; BPI, Brief Pain Inventory; ODI, Oswestry Disability Index; DDS, 30 Descriptor Differential Scale; SIP, Sickness Impact Profile; WOMAC, Western Ontario and McMaster Universities Osteoarthritis Index; MPQ-SF, McGill Pain Questionnaire Short-Form; PDI, Pain Disability Index; DN4, 31 Douleur Neuropathique 4; MADRS, Montgomery–Åsberg Depression Rating Scale; GDS, Geriatric Depression Scale; HDS, Hospital Depression Scale; SNRI, Serotonin–norepinephrine reuptake inhibitors; SSRI, Selective 32 serotonin reuptake inhibitors. 33 34 35 36 37 38 39 40 41 42 35 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 56 BMJ

1 2 3 Supplemental file 10. Risk of bias of included trials 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

36 https://mc.manuscriptcentral.com/bmj BMJ Page 56 of 56

1 2 3 Supplemental file 11. Exploratory meta-regression analyses of risk of bias, small study effects, industry funding, 4 5 depression diagnosis and daily dosage antidepressants on pain. P-value for 6 Moderator Weighted mean difference (95% CI) I2 7 interaction 8 RiskConfidential: of bias For Review Only 9 High (25 trials) -7.6 (-11.0. to – 4.3) 86.5% 0.56 10 Low (5 trials) -9.8 (-13.9 to -5.7) 0% 11 Small study 12 Yes (20 trials) -6.7 (-10.3 to -3.2) 55.1% 0.49 13 No (10 trials) -9.0 (-13.6 to -4.4) 92.9% 14 Industry funding 15 16 Yes (14 trials) -6.9 (-9.0 to -4.8) 53.8% 17 No (11 trials) -7.5 (-15.2 to 0.3) 91.1% 0.24 18 Unclear (5 trials) -13.3 (-21.9 to -4.7) 55.1% 19 Depression diagnosis 20 Yes (4 trials) -8.4 (-16.3 to -0.5) 54.4% 0.25 21 No (26 trials) -7.8 (-11.0 to -4.6) 85.8% 22 Daily dosage 23 duloxetine 20 mg/day (1 trial) 0.8 (-7.2 to 8.8) 0% 24 duloxetine 60 mg/day (9 trials) -7.6 (-10.1 to -5.2) 59.4% 0.13 25 duloxetine 120 mg/day (4 trials) -9.2 (-11.6 to -6.8) 0% 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 37 43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60