Comprehensive Review

Cannabinoids, cannabis, and cannabis-based medicines for pain management: an overview of systematic reviews

Moore, R. Andrewa,*; Fisher, Emmab,c; Finn, David P.d; Finnerup, Nanna B.e,f; Gilron, Iang,h,i; Haroutounian, Simonj,k; Krane, Elliotl; Rice, Andrew S.C.m; Rowbotham, Michaeln,o; Wallace, Markp; Eccleston, Christopherb,c,q

Author Information

aNewton Ferrers, Plymouth, United Kingdom

bCentre for Pain Research, University of Bath, Bath, United Kingdom

cCochrane Pain, Palliative, and Supportive Care Review Groups, Oxford University Hospitals, Oxford, United Kingdom

dPharmacology and Therapeutics, School of Medicine, Galway Neuroscience Centre and Centre for Pain Research, NCBES, National University of Ireland Galway, Galway, Ireland

eDepartment of Clinical Medicine, Danish Pain Research Center, Aarhus University, Aarhus, Denmark

fDepartment of Neurology, Aarhus University Hospital, Aarhus, Denmark

gDepartment of Anesthesiology and Perioperative Medicine, Kingston General Hospital and Queen's University, Kingston, ON, Canada

hCentre for Neuroscience Studies, Queen's University, Kingston, ON, Canada

iSchool of Policy Studies, Queen's University, Kingston, ON, Canada

jDivision of Clinical and Translational Research, Washington University Pain Center, St. Louis, MO, United States

kDepartment of Anesthesiology, Washington University School of Medicine, St Louis, MO, United States

lDepartment of Anesthesiology, Perioperative and Pain Medicine, and Pediatrics, Stanford University School of Medicine, Stanford, Palo Alto, CA, United States

mPain Research, Department of Surgery and Cancer, Faculty of Medicine, Imperial College London, United Kingdom

nDepartment of Anesthesia, Pain Management Center, University of California San Francisco, San Francisco, CA, United Kingdom

oSutter Health, CPMC Research Institute, California Pacific Medical Center Research Institute, San Francisco, CA, United States

pDivision of Pain Medicine, Department of Anesthesiology, University of California San Diego, San Diego, CA, United States

qDepartment of Clinical and Health Psychology, Ghent University, Ghent, Belgium

*Corresponding author. Address: Appledore, 84 Court Road, Newton Ferrers, Plymouth PL8 1DA, United Kingdom. Tel.: +44 1752 873532. E-mail address: [email protected] (R.A. Moore).

Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article.

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Web site (www.painjournalonline.com).

PAIN 162():p S67-S79, July 2021. | DOI: 10.1097/j.pain.0000000000001941

Abstract

Cannabinoids, cannabis, and cannabis-based medicines (CBM) are increasingly used to manage pain, with limited understanding of their efficacy and safety. We assessed methodological quality, scope, and results of systematic reviews of randomised controlled trials of these treatments. Several search strategies sought self-declared systematic reviews. Methodological quality was assessed using both AMSTAR-2 and techniques important for bias reduction in pain studies. Of the 106 articles read, 57 were self-declared systematic reviews, most published since 2010. They included any type of cannabinoid, cannabis, or CBM, at any dose, however administered, in a broad range of pain conditions. No review examined the effects of a particular cannabinoid, at a particular dose, using a particular route of administration, for a particular pain condition, reporting a particular analgesic outcome. Confidence in the results in the systematic reviews using AMSTAR-2 definitions was critically low (41), low (8), moderate (6), or high (2). Few used criteria important for bias reduction in pain. Cochrane reviews typically provided higher confidence; all industry-conflicted reviews provided critically low confidence. Meta-analyses typically pooled widely disparate studies, and, where assessable, were subject to potential publication bias. Systematic reviews with positive or negative recommendation for use of cannabinoids, cannabis, or CBM in pain typically rated critically low or low (24/25 [96%] positive; 10/12 [83%] negative). Current reviews are mostly lacking in quality and cannot provide a basis for decision-making. A new high-quality systematic review of randomised controlled trials is needed to critically assess the clinical evidence for cannabinoids, cannabis, or CBM in pain.

1. Introduction

In 2018, the International Association for Study of Pain established a Task Force on the use of cannabinoids, cannabis, and cannabis-based medicines (CBM) for pain management. It has 4 Work Packages (WP) focused on (1) basic science and evidence for efficacy in preclinical models,72 (2) evidence for clinical analgesic efficacy,28 (3) risk and evidence of harm to the individual,31,50 and (4) the societal impact and policy.

This overview review is part of the second Work Packages and is focused on summarizing the evidence of efficacy presented in systematic reviews of randomised controlled trials (RCTs) of any broadly defined cannabinoid product in any type of pain condition.

There have long been concerns about the quality of most of the medical literature.37,68 A 1996 survey indicated that 90% of meta-analyses had methodological flaws that could limit their validity, and that meta-analyses of low quality produced significantly more positive conclusions.39 There has subsequently been an epidemic of systematic reviews, with huge growth rates in their numbers without any necessary improvement in their quality,68 leading to the conclusion that the large majority of systematic reviews and meta-analyses are unnecessary, misleading, and/or conflicted. Industry-supported meta-analyses have been found to be less transparent, with few reservations about methodological limitations of the included trials, and with more favourable conclusions than corresponding Cochrane reviews.41

Even in good-quality Cochrane reviews, the use of Grading of Recommendations Assessment, Development and Evaluation (GRADE) to summarise the quality of evidence indicates that fewer than 20% of reviews actually have any high-quality evidence.30 Only confusion results from the product of low-quality evidence and low-quality systematic review of that evidence. High standards in clinical trials and systematic reviews of those trials are absolutely essential to improve knowledge, make policy, or make individual clinical decisions; anything else is guesswork.

The aim of this overview review of WP2 was therefore to assess the methodological quality, scope, and reported results of systematic reviews and meta-analyses of RCTs of cannabinoids, cannabis, and CBM for pain relief, and to determine whether any new systematic review was required.

2. Methods

A protocol for this overview is registered on Prospero (Prospero ID CRD42019124710) and published.28 We followed the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols.49 A truncated report of methods is therefore given here.

2.1. Systematic reviews for inclusion

We chose to include any review whose authors defined as a systematic review of RCTs, although definitions of what comprises a systematic review are considerably more restrictive.34 The intention in the protocol was to analyse only high-quality evidence found in Cochrane Effective Practice and Organisation of Care25 GRADEs of evidence as high or moderate quality, although in the event this was not possible.

2.2. Participants

We included systematic reviews of RCTs involving people of any age with any form of acute and chronic pain, including pain secondary to another condition, such as pain with spasm, multiple sclerosis (MS), or leg cramps. Experimental pain was not included.

2.3. Interventions and comparators

We included any type of cannabinoid product (natural or synthetic), cannabis, or CBM, by any route of administration, at any dose, and with any comparison intervention. This included endocannabinoid system modulators such as fatty acid amide hydrolase inhibitors and N-palmitoylethanolamide. For inclusion, systematic reviews were required to examine interventions to reduce pain intensity.

2.4. Outcomes

A range of primary and secondary outcomes were proposed in the protocol. The limited nature of systematic reviews included meant that this overview was restricted to measures of analgesic efficacy only.

2.5. Search and selection of systematic reviews

We searched PubMed, EMBASE, DARE, and the Cochrane Controlled Register of Trials (CENTRAL) for systematic reviews of cannabinoids, cannabis, and CBM and for people with pain. The main search was completed in August 2019. Bibliographies of included and excluded reviews were examined for possible reviews, and electronic citations of PubMed and Google Scholar also examined. We conducted targeted searches for further systematic reviews through additional electronic searches, through reference lists of retrieved articles and reviews, or through other sources up to January 2020. Two authors independently sifted the titles and abstracts identified, with a third author resolving disagreement.

2.6. Data extraction

Data extraction was conducted by 2 authors (R.A.M. & E.F.) and checked by others; disagreement was resolved by consensus (initially discussion with C.E., but wider if needed). The following information was extracted from each review:

  • (1) Review characteristics, eg, design, participants, age and sex, pain condition, inclusion/exclusion criteria, and risk of bias method. In addition, we sought information related to Assessment of Multiple Systematic Reviews (AMSTAR)-2,67 methodological issues related to known sources of bias in pain studies, use of GRADE, and GRADE assessment if used. We extracted the number of trials and patients used in assessment of pain, the number in randomised trials, and the number in randomised trials used to calculate any summary estimate of analgesic efficacy.
  • (2) Intervention and comparator characteristics, eg, type of cannabinoid, dose, route of administration, and comparator.
  • (3) Outcomes listed in the primary outcomes.

In addition, we made a judgement on the strength of any recommendation made by authors in the review abstract. This was done by R.A.M. and E.F., discussed initially with C.E., and then wider if needed. Judgements were based on being positive (recommended use), equipoise (a balance of evidence), or negative (recommended nonuse), or no statement. Strength was judged as strong (eg, provide reasonable therapeutic option, should not be used), moderate (eg, moderately effective, no unbiased evidence), or weak (eg, small analgesic benefit, more trials needed to support).

2.7. Assessment of review quality and validity

We assessed each included review using AMSTAR-2.67 Two authors assessed each review using the criteria with disagreement resolved by consensus.

We also conducted additional validity checks of potential critical importance in the evaluation of analgesic efficacy. These included:

  • (1) Did the review use a defined diagnostic criterion for pain conditions?
  • (2) Did the reviews include only studies in which patients made their own assessment of pain? (professional and patient assessment often disagrees, with professionals significantly underestimating pain66).
  • (3) Did the reviews use studies with defined minimum pain intensity of moderate or severe pain? (mild pain can reduce the sensitivity of trials to demonstrate an analgesic effect).
  • (4) Did the reviews examine study size as a confounding factor in any analysis of efficacy? (systematic reviews have been criticised for being overconfident of results with inadequate data2,65,79; there is increasing evidence of the importance of small trial size, both because of random chance,16,52,76 and as an important source of bias.22,23,26,36,57
  • (5) Did the review examine susceptibility to publication bias? (if possible, for each review with dichotomous numerical data, we will assess the likelihood of publication bias51).

2.8. Data synthesis and/or descriptive evaluation of reviews

We planned to evaluate strengths and weaknesses of systematic reviews of cannabinoids in treating pain by a descriptive analysis, and meta-analysis if appropriate. Because there was little high-quality evidence found, and none useable for meta-analysis, the descriptive analysis of all systematic reviews became the only mechanism of evaluation.

The absence of combinable data did not preclude unplanned evaluations, including links between AMSTAR-2 score-matched critical methodological criteria for evaluation of analgesic efficacy, the potential impact of reviews using Google Scholar citation numbers, judgement on the strength of recommendations by review authors based on language in the abstracts, numerical assessment of analgesic efficacy, and how these factors interacted.

The nature of the reviews precluded any useful independent assessment of GRADE for individual reviews.

3. Results

3.1. Results of search

Initial electronic searching found 685 possible systematic reviews, reduced to 559 after removal of duplicates. After reading the abstracts, 106 articles were obtained in full, and read for possible inclusion as a systematic review. Forty-nine were excluded (Appendix 1 lists the reasons and references, available as supplemental digital content at https://links.lww.com/PAIN/B61). Reasons for exclusions were:

  • (1) 12 were narrative or scoping reviews;
  • (2) 11 were practice guidelines, position papers, or therapeutic recommendations;
  • (3) 8 were overview reviews;
  • (4) 2 each were not systematic reviews, had no pain outcomes; did not investigate cannabinoids, were duplicates, were older versions of updated systematic reviews, or were thesis chapters not relevant to this overview;
  • (5) 1 each was a clinical trial, a review of consensus statements, investigated placebo only, investigated experimental pain, investigated adverse events, or investigated combination therapies without cannabinoids alone.

Fifty-seven3,5–15,17,19–21,24,27,29,32,33,35,38,40,42–48,54–56,58–64,69–71,73–75,77,78,80–87 were included in this overview review as systematic reviews investigating cannabinoids, cannabis, or CBM in RCTS with pain as an outcome (Appendix 2 gives information on date of publication, pain or other condition examined, type of CBM investigated, route of administration, and abstract conclusion; available as supplemental digital content at https://links.lww.com/PAIN/B61). Seven were Cochrane reviews,5,10,15,54,59,64,82 and 6 reported either financial sponsorship from a pharmaceutical company or were conflicted because authors were employees of a company with interests in cannabinoids.8,38,46,69–71

Forty-seven reviews specifically examined pain (10 neuropathic pain, 10 any type of pain, 8 chronic noncancer pain, 7 cancer pain, 4 rheumatic pain, 2 fibromyalgia, 2 spinal cord injury, 1 acute pain, 1 pain in children, 1 HIV neuropathy, and 1 phantom limb pain). Ten reviews considered conditions in which pain was one of several symptoms (5 muscular dystrophy, 3 neurological conditions, 1 leg cramps, and 1 gastrointestinal conditions).

3.2. General description of included studies

Included studies were mostly published recently. The earliest was published in 200117 and the most recent in 2019.13,33,56,63,86 Only 5 were published before 2010; most (90%) were published since 2010 (Fig. 1), and two-thirds of the included reviews had been published in the past 5 years.

F1
Figure 1.:
Rate of publication of 54 included systematic reviews. Number of systematic reviews on cannabinoids, cannabis, or CBM published each year between 2001 and 2019. No systematic review was published before 2001. CBM, cannabis-based medicines.

Summary information on these 57 systematic reviews is shown in Table 1, organised by pain condition. Because several of these systematic reviews examined a wide range of medical conditions, only the information involving pain and cannabinoids is shown.

Table 1 - Summary information on included reviews, with data for pain and cannabinoid only.
Type of pain Number of reviews Total for pain RCT Analysis
Studies Patients Studies Patients Studies Patients
Acute pain 1 7 611 7 611 None None
Chronic noncancer 8 6–91 226–N/A 6–47 226–N/A 0–9 0–1734
Fibromyalgia 2 2 69–72 2 69–72 None None
GI 1 1 21 1 21 None None
HIV 1 2 89 2 89 2 89
Leg cramps 1 1 22 1 22 1 22
MS 5 3–11 N/A–2692 3–11 N/A–2692 0–11 0–2692
Neurology (other) 3 0–7 0–17 0–7 0–17 None None
Neuropathic 10 0–16 0–1750 0–16 0–1750 0–10 0–1001
Phantom 1 0 0 0 0 None None
Rheumatic 4 0–4 0–203 0–4 0–203 1 58
SCI only 2 0–2 0–20 0–2 0–20 None None
Cancer only 7 0–8 0–1539 0–8 0–1539 2–5 537–1333
All pain 10 4–43 N/A–2454 4–43 N/A–2437 0–25 0–2248
Children (all pain) 1 4 19 0 0 None None
MS, multiple sclerosis; N/A, not available in the review; RCT, randomised controlled study.

Neuropathic pain, chronic noncancer pain, all pain, and cancer pain predominated, but 15 distinct areas were recognised. Some were not classic pain conditions, for example, pain associated with spasm in MS or leg cramps.

The number of studies and patients involved, the number of randomised trials included, and the number available for any statistical analysis varied widely between reviews and pain conditions. For 6 pain conditions, there were no available data for analysis. Some reviews in some conditions analysed data from several thousand patients. For example, Yanes et al.86 analysed data from 25 trials and 2248 patients in a meta-regression for all pain conditions, and Torres-Moreno et al.77 pooled data from 2692 patients in 11 MS trials. The variation in numbers between reviews in apparently the same condition reflected different approaches in analysing drug and formulation, route of administration, and whether a broad or narrow approach to type of condition was used (eg, painful diabetic neuropathy rather than all neuropathic pain).

Risk of bias was assessed in 45 of the 57 reviews, predominantly using Cochrane risk of bias tools (27), the Oxford quality scale (13) or a modification of it (3), an American Academy of Neurology tool (1), or a Physiotherapy Evidence Database tool (1). Twelve systematic reviews made no apparent mention of formal risk of bias assessment.8,11,12,20,42,60,62,78,80,85–87

3.3. Cannabinoid used and route of administration

Reviews examined the effects of a wide range of cannabinoid drugs or preparations, although these were seldom clearly defined. The most common definition was “cannabinoid” in 36 reviews, “any cannabis preparation” in 8, “plant-based cannabis preparation” in 6, 2 each examined nabilone, dronabinol, or nabiximols, 2 palmitoylethanolamide derivatives, and one each Δ9-tetrahydrocannabinol (THC), cannabidiol, and Cannabis sativa. Appendix 3 shows their use by pain condition (available as supplemental digital content at https://links.lww.com/PAIN/B61).

Route of administration was generally not defined (36 reviews), or “any route” (15 reviews), oral or topical (4 reviews), or smoked or inhaled (2 reviews) (Appendix 4, available as supplemental digital content at https://links.lww.com/PAIN/B61).

3.4. Pain outcome used in systematic reviews

Few reviews clearly defined an outcome of at least 30% or 50% pain intensity reduction considered by patients to be a good outcome53(see Appendix 5 for results according to pain condition, available as supplemental digital content at https://links.lww.com/PAIN/B61); only 14 reviews (25%) sought these outcomes. Most, 28 reviews, did not define an outcome. The remaining reviews used some calculation based on continuous variables, mostly standardised mean difference or mean difference (10 reviews), effect size (3 reviews), or other (2 reviews). Three-quarters of the reviews therefore made no prior adjudication of what a successful outcome might be or justified the choice of outcome.

3.5. Use of GRADE

GRADE evaluation of overall evidence certainty was used in 17 systematic reviews, and not used in 40. Of the 17 that used GRADE, 13/17 (76%) included very low certainty; the ratings were:

  • (1) 0 high certainty
  • (2) 4 moderate certainty
  • (3) 1 moderate or very low certainty
  • (4) 3 low certainty
  • (5) 2 low or very low certainty
  • (6) 7 very low certainty.

3.6. Assessment of multiple systematic reviews rating

Each review was judged according to the 16-criterion AMSTAR 2 list; Appendix 6 shows scoring for individual reviews (available as supplemental digital content at https://links.lww.com/PAIN/B61). The results (Table 2) showed that only 4 of the criteria were met by more than half of the reviews (showing some detail of included studies [85%], conflict of interest reporting [81%], study selection in duplicate [65%], and data extraction in duplicate [56%]).

Table 2 - The extent to which reviews met individual AMSTAR 2 criteria (critical criteria in italic and noncritical criteria in bold).
Number Criterion Met Not met Partially met Not applicable
1 PICO 13 44
2 Protocol established before review 17 40
3 Study design explanation 0 57
4 Comprehensive literature search 1 3 53
5 Study selection in duplicate 36 21
6 Data extraction in duplicate 33 24
7 List of exclusions + justification 26 31
8 Included studies in detail 48 9
9 RoB assessment satisfactory 27 13 17
10 Sources of funding for studies 10 47
11 Appropriate meta/analysis 20 8 29
12 Impact of RoB on result 17 25 15
13 Impact of RoB in discussion/interpretation 22 31 4
14 Heterogeneity investigated discussed 22 28 7
15 Publication bias (small study bias) 14 36 7
16 CoI of review authors reported 45 12
Bars indicate critical AMSTAR-2 criteria.
CoI, conflict of interest; PICO, patient, intervention, comparator, outcome; RoB, risk of bias.

Meeting some criteria was not applicable. For example, because many reviews had few studies, with those studies often small and clinically heterogeneous, no meta-analysis was appropriate. That situation made it difficult to discuss risk of bias or heterogeneity on the result. It was also difficult to judge whether risk of bias was satisfactory. In addition, the AMSTAR criterion for a comprehensive literature search demanded searches of gray literature, consulting experts in the field, and searching of trial registries. One review met all those criteria,13 and 53 reviews partially met the criteria for a literature search, typically not searching gray literature.

Other criteria met uncommonly were a study design explanation (no reviews), reporting on sources of funding of studies (19 reviews), production of a complete PICO (Patient, Intervention, Comparison, Outcome; 13 reviews), assessment of small study bias (14 reviews), and prior publication of a protocol (17 reviews).

The failure to meet critical and noncritical AMSTAR criteria resulted in generally low assessments of overall confidence in the results of the review. Using the AMSTAR definitions, confidence in the results of 86% of the reviews was critically low (41 reviews) or low (8 reviews). For 6 reviews, confidence in the results was moderate, and for only 2 was it high; only 1 in 7 systematic reviews had moderate or high AMSTAR confidence.

For the 7 Cochrane reviews, overall confidence was high or moderate in 4, and low in 3; none was critically low. All 6 conflicted reviews were assessed as having critically low confidence (Fig. 2). Of the 12 reviews not specific for pain, 9 were evaluated as critically low, and 3 of low confidence.

F2
Figure 2.:
AMSTAR evaluation of confidence in results of systematic reviews. AMSTAR-2 evaluations as percentages of 57 systematic reviews, 7 Cochrane reviews, and 6 conflicted reviews with industry sponsorship.

Appendix 7 shows AMSTAR confidence assessments in the results of the reviews according to the pain condition investigated (available as supplemental digital content at https://links.lww.com/PAIN/B61).

3.7. Use of critical criteria for assessment of analgesic efficacy

Because AMSTAR is a generic instrument not specifically designed for use with analgesic studies, we also assessed the systematic reviews for the use of methodologies designed to avoid significant biases in pain trials (Appendix 6 for individual result scoring and Appendix 8 for summary results by pain condition; available as supplemental digital content at https://links.lww.com/PAIN/B61). The results for this assessment demonstrated that methods used to minimise bias in analgesic studies were typically not used (Fig. 3).

F3
Figure 3.:
Percentage use of criteria designed to minimise bias in analgesic trials. Percentage of reviews using criteria designed to minimise bias in analgesic trials. BOCF, baseline observation carried forward; LOCF, last observation carried forward.

Only half of the reviews specified using trials that were both properly randomised and properly double blind. About 10% or fewer required patient-reported pain only, performed any sensitivity analysis for small trials, evaluated sensitivity of a result to publication bias, used studies where entrants were required to have pain with a minimum pain intensity, or evaluated the potential impact of imputation method for missing data.

Results for the 7 Cochrane reviews overall confidence were mixed. Three from a pain-specific group used only randomised and double-blind studies and met 50% or more of the criteria. The other 4 accepted studies were not necessarily randomised or double blind and met 25% of criteria or fewer. Four of the 6 conflicted reviews used only randomised and double-blind studies and met 38% or fewer of the criteria. Of the 12 reviews not specific for pain, only 4 used randomised and double-blind studies and 11 met 25% or fewer of the criteria.

3.8. Concordance between assessment of multiple systematic reviews and critical pain criteria

Thirty-seven of the 39 (95%) reviews with AMSTAR ratings of critically low met 2 or fewer of the 8 critical pain criteria compared with 5/8 (63%) rated low, and 2/7 (29%) of those rated moderate or high. There was a tendency for higher scores in reviews with AMSTAR rating of moderate or high (Table 3).

Table 3 - Concordance between AMSTAR rating and critical pain criteria score (of 8).
AMSTAR rating Range Mean Median
Critically low 0–6 1.5 2.0
Low 1–3 2.1 2.0
Moderate or high 1–6 4.0 4.0

A few individual reviews stood out against the trend. For example, Finnerup et al.27 had a critically low AMSTAR rating, but met 6 of 8 critical pain criteria, whereas Brettshneider et al.15 had a high AMSTAR rating but scored only 1 of 8 criteria. Results of the 3 Cochrane reviews from a pain-specific group scored reasonably well on both.

3.9. Impact of systematic reviews

We examined the impact of the 54 of the 57 systematic reviews by looking at the number of citations for each as reported by Google Scholar on Sept 20 2019; three 2019 publications were identified after that date, and would have had few if any citations. The 54 reviews had been cited a total of 6760 times (mean 125 citations, median 47 citations), with about 1100 citations a year (mean 22 citations a year, median 9 citations a year).

The majority of citations came for a small number of systematic reviews (Fig. 4). Two reviews were cited on average over 200 times a year,27,84 and 3 other reviews were cited between 50 and 100 times a year.5,42,54 Because most reviews had been published within a limited period, no effect of recency in citation rate was discernible.

F4
Figure 4.:
Distribution of total citations and total publications according to annual rate of citation. Distribution of the percentage of total citations and total reviews according to the average number of citations per year for each review calculated from the number of Google Scholar citations divided by the number of years since publication. A small percentage of reviews accounted for the largest percentage of citations.

There was no consistency as why these 5 were particularly heavily cited. Table 4 summarises information on these reviews, with possible reasons for their high citation rate.

Table 4 - Information on 5 systematic reviews with high average annual citation rates.
Review and journal Average annual citations AMSTAR rating Pain criteria met Probable reason for high citation rate
Finnerup,27 Lancet Neurology 320 Critically low 6/8 Broad systematic review of interventions for neuropathic pain from a Special Interest Group of International Association for the Study of Pain published in Lancet Neurology.
Whiting et al.,84 JAMA 209 Critically low 2/8 Systematic review of cannabinoids for medical use published in JAMA—pain was only one component, and highlighted in the section on chronic pain of the 2017 national Academies report in the United States.
Koppel et al.,42 Neurology 70 Critically low 1/8 Report of a guideline development committee of the American Academy of Neurology and published in Neurology.
Mücke et al.,55 Cochrane 56 High 6/8 Broad review of pharmacologic interventions for treating phantom limb pain published in Cochrane.
Alviar et al.,5 Cochrane 50 Moderate 4/8 Systematic review of cannabis-based medicines for neuropathic pain published in Cochrane.

3.10. Judgement on abstract strength of recommendation

Abstracts of systematic reviews typically provide some comment on the strength of evidence or the direction of any effect, and frequently both. These can be interpreted as recommendations or are frequently framed as recommendations. We made a judgement as to whether the implied recommendation in an abstract was positive or negative (strong, moderate, or weak), or showed equipoise, or whether there was no recommendation relating to the use of cannabinoid-based medicines for treatment of pain. Table 5 shows examples of how we made these judgements, with details of each in Appendix 7 (available as supplemental digital content at https://links.lww.com/PAIN/B61).

Table 5 - Examples of judgements on abstract strength of recommendation.
Judgement on abstract strength of recommendation Number Examples
No opinion 13 No recommendation relating to CBM and pain
Strong positive 6 “Currently available cannabinoids are safe, modestly effective analgesics that provide a reasonable therapeutic option in the management of chronic noncancer pain” (Lynch 2015)
“Cannabinoids including the cannabidiol/THC buccal spray are effective in treating neuropathic pain in MS” (Iskedihan 2007)
“…Cannabinoid-based pharmacotherapies may serve as effective replacement/adjunctive options regarding pain” (Yanes 2019)
Moderate positive 12 “There is evidence that cannabinoids are safe and modestly effective in neuropathic pain with preliminary evidence of efficacy in fibromyalgia and rheumatoid arthritis” (Lynch 2011)
“Cannabis and cannabinoids provide an interesting treatment choice for PDN” (Alessa 2018)
Weak positive 10 “Limited evidence suggests that cannabis may alleviate neuropathic pain in some patients” (Nugent 2017)
“Selective cannabinoids provide a small analgesic benefit in patients with chronic neuropathic pain” (Meng 2017)
Equipoise 3 “The quality of currently available evidence on the effectiveness of adjuvant analgesics in the treatment of cancer pain is low” (van den Beuken 2017)
Weak negative 6 “More trials with rigorous design and reporting are needed” (Jawahar 2013)
“Very low-quality evidence suggests that oromucosal nabiximols and THC have no effect on pain” (Häuser 2019)
Moderate negative 3 “No convincing, unbiased, high-quality evidence suggesting that nabilone is of value in treating people with fibromyalgia” (Walitt 2016)
“Smoked cannabis cannot be recommended as routine therapy” (Phillips 2010)
Strong negative 4 “Cannabinoids have no role in the management of acute pain” (Stevens 2017)
“The available evidence does not support current use in the management of rheumatoid arthritis” (Macfarlane 2011)
“Do not offer [nabilone, dronabinol, THC, a combination of cannabidiol (CBD) with THC] to manage chronic pain in adults” (NICE 2019)
“…For adults with advanced cancer, the addition of cannabinoids to opioids did not reduce cancer pain.” (Boland 2019)
CBM, cannabis-based medicines.

Of the 44 systematic reviews making a recommendation, 27 were moderate or strong, 10 weak, and 7 indicated equipoise. Of the 7 Cochrane reviews, 3 had no opinion, 2 were strong negative, one at equipoise, and one was weak positive. Of the 6 conflicted reviews, 3 had no opinion, one was weak positive, and 2 were strong positive. Eight of the 19 reviews with moderate or strong positive recommendation had no formal risk of bias assessment.8,11,12,20,42,60,61,86 All reviews making weak recommendations had a risk of bias assessment.

3.11. Relationship between abstract strength of recommendation, assessment of multiple systematic reviews rating, and critical pain criteria

Table 6 shows the AMSTAR rating and critical pain criteria score. Systematic reviews with any positive recommendation of use of cannabinoids, cannabis, or CBM in pain typically rated critically low on AMSTAR (20/25; 80%) or critically low or low (24/25; 96%). They typically used 2 or fewer of the 8 critical pain criteria to avoid bias. Reviews with negative recommendations were somewhat less likely to have a critically low (6/12; 50%) or critically low or low rating (10/12; 83%), and used 3 or more critical pain criteria.

Table 6 - Relationship between abstract strength of recommendation and AMSTAR rating and critical pain criteria.
No opinion Positive Equipoise Negative
Any Strong or moderate Weak Weak Strong or moderate Any
Number 13 25 19 6 7 4 8 12
AMSTAR rating
 Critically low 10 20 17 3 5 2 4 6
 Low 1 4 1 3 0 1 2 3
 Moderate 1 1 1 0 1 1 2 3
 High 1 0 0 0 1 0 0 0
Critical pain criteria score (of 8)
 Mean 1.6 1.5 1.4 1.7 2 4.0 2.8 3.2
 Median 1 2 2 1.5 2 4 2 3
 Range 0–4 0–5 0–5 1–3 1–3 2–6 1–4 1–6

Reviews making weak positive or negative recommendations were more likely to use more critical pain criteria than those making moderate or strong recommendations.

3.12. Calculation of magnitude of analgesic effect

Seventeen systematic reviews used numeric data to calculate the magnitude of any analgesic effect, for a variety of interventions, routes of administration, and painful condition, using a range of different pain outcomes and statistical outputs (Table 7). Analyses were conducted on data sets that varied between 1 and 25 trials, and 22 and 2692 patients. No analysis was conducted on a defined intervention, defined dose or intensity, or defined route of administration in a defined pain condition.

Table 7 - Magnitude of analgesic effect calculations in 15 systematic reviews reporting it.
Review Intervention Route of administration Condition Trials/patients in analysis Outcome Statistical result Strength of abstract recommendation AMSTAR-2 confidence rating Pain criteria (of 8 points)
Andreae et al.7 Cannabis sativa Inhaled Neuropathic 5/185 ≥30 pain reduction NNT 5.6 (3.4 to 13) Strong + Moderate 5
Yanes et al.86 Cannabinoids Any All pain 25/2248 Effect size, mean Cohen's d = −0.58, 95% CI (−0.74 to −0.43) Strong + Critically low 0
Artukoglu et al.8 Palmitoylethanolamide Not defined All pain 8/1244 Unclear WMD 2.03 (1.19 to 2.87) Moderate + Critically low 0
Martin-Sanchez et al.46 Any cannabis preparation Not defined Chronic noncancer 7/288 Effect size, mean −0.61 (−0.84 to −0.37) Moderate + Critically low 2
Nugent et al.58 Plant-based cannabis preparations Not defined Chronic noncancer 9/1042 neuropathic only ≥30 pain reduction RR 1.4 (1.2 to 1.9) Moderate + Critically low 2
Whiting et al.84 Cannabinoids Not defined All pain 8/1370 ≥30% pain reduction OR 1.41 (0.99 to 2.00) Moderate + Critically low 2
Aviram et al.9 Any cannabis-based medicines Any All pain 24/1334 Mean change 0-10 scale −0.61 (−0.78 to -0.43) Weak + Critically low 2
Meng et al.48 Dronabinol, nabilone, naboximols Any Neuropathic 10/973 Mean difference 0-10 scale −0.65 (−1.06 to −0.23) Weak + Critically low 3
Torres-Moreno et al.77 Cannabinoids Oral or oromucosal MS pain 11/2692 SMD −0.17 (–0.31 to –0.03) Weak + Critically low 2
da Rovere et al.19 Cannabinoids Not defined MS pain 3/not stated SMD 0.55 (−3.3 to 4.4) None Critically low 1
Jawahar et al.40 Cannabinoids Not defined MS pain 3/565 Effect size, mean 0.08 (−0.74 to 0.89) None Critically low 3
Stocking et al.74 Cannabis and cannabinoids Any Chronic noncancer 5/753 ≥50% pain reduction OR 1.4 (0.97 to 2.1) None Moderate 3
8/1734 ≥30% pain reduction OR 1.4 (1.2 to 1.8) None Moderate 3
Finnerup et al.27 Cannabinoids Any Neuropathic 5/815 ≥50% reduction RD 0.027 (−0.034 to 0.088) Weak − Critically low 6
Mucke 201854 Any cannabis-based medicines Any Neuropathic 8/1001 ≥50% reduction RD 0.05 (0.00 to 0.09) Weak − High 6
Boland 201913 Cannabinoids Any Cancer 5/1442 Mean difference 0-10 scale Mean difference −0.21 (−0.48 to 0.07, P = 0.14) Strong − Moderate 3
Baldinger et al.10 THC Not defined Muscle cramps 1/22 VAS, mean 0.24 (−0.31 to 0.79) Strong − Low 2
Phillips et al.61 Any Not defined HIV neuropathy 2/89 ≥30 pain reduction RR 2.4 (1.4 to 4.1) Strong − Critically low 3
Bold entries were significantly different from placebo. NNT, number-needed-to-treat.

Ten of 18 reported results showed statistical difference from placebo, whereas 8 showed no difference. Statistical significance was associated with positive abstract recommendation, whereas no significance was associated with negative abstract recommendations.

Table 7 also shows AMSTAR-2 confidence level and the number of points scored for critical pain elements. Statistical significance was generally associated with critically low AMSTAR assessment and 3/8 or fewer pain criteria. The exception was a review of short-term (6 hours to 5 days) effects of inhaled or smoked cannabis in small number of patients with chronic pain conditions in trials with significant risk of bias.7

Reviews without statistical significance generally made no recommendation or a negative recommendation, with a tendency to higher AMSTAR grades and higher scores for critical pain elements. Exceptions were the study by Stockings et al.74 with bare significance for one outcome, and the study by Phillips et al.,61 which included only 89 patients.

A calculation of the susceptibility to publication bias from studies with null results was based on a number-needed-to-treat of 10 or greater being clinically not relevant.51 No calculation was possible where there was no statistical significance (number-needed-to-treat is then essentially infinity), or for reviews reporting continuous outcomes. For only 3 of these reviews were data available that allowed for the calculation. The number of people in trials with null results required to overturn the statistically significant results for cannabinoids was 213 in Nugent et al.,58 237 in Phillips et al.,61 and 145 in Andreae et al.7

4. Discussion

To the extent that any conclusions can be drawn from existing systematic reviews, they can only be made with respect to the types of cannabinoid, cannabis, and CBM investigated to date, in the specific patient groups and pain types studied. No conclusions with respect to other interventions yet to be tested are possible.

Of the many reviews of the effects on pain of studies concerning cannabinoids, cannabis, and CBM published to the end of 2019, 57 claimed to be systematic reviews. The Cochrane Handbook lists 5 key characteristics of a systematic review34: a clearly stated set of objectives, explicit and reproducible methods, systematic and comprehensive searches, assessment of validity of results including risk of bias, and a systematic presentation of those results. Although relatively few of the 57 systematic reviews examined here would meet these criteria, the cachet of a review claiming to be systematic is such that all required examination.

Included reviews examined 15 distinct pain areas. Most accepted any type of cannabinoid, at any dose, by any route of administration, although some had a more defined scope. No review examined the effects of a particular cannabinoid, at a particular dose or dose range, given by a particular route of administration, for a particular pain condition, and reporting a particular analgesic outcome.

The degree of pooling for any effect calculations could be extreme; for example, one review of cannabinoids in chronic pain pooled data from trials as short as 6 hours, and as long as 15 weeks.57 Almost 80% of the reviews failed to define what a successful outcome might be; rather than defining a measure that is important to patients,54 the approach was to make important that which had been measured.

Judging systematic review quality is difficult. The most often used methodology is AMSTAR, and we used the most recent version, AMSTAR-2.67 This is a generic tool examining what is regarded as best practice in systematic review methodology. The 16 items are themselves not controversial, although judging whether a review meets a particular criterion is often subjective, open to disagreement between assessors because details of methods may be omitted from publications for reason of space. An AMSTAR-2 score probably represents a “worst case.” No single criterion was met by more than 50% of the reviews, and confidence in the results of 86% of the reviews was critically low (41 reviews) or low (8 reviews). This is similar to an assessment of systematic reviews in back pain where 74% rated critically low and 16% low.4

A generic approach to quality is of limited value in assessing analgesic effects because it does not examine criteria known to be associated with considerable bias in pain trials. We examined 8 of those with established association with risk of bias, and again few reviews used them. Randomisation and blinding were defined criteria in just over half of the reviews, but 5 of the 8 criteria were used by 10% or fewer of the reviews. Reviews with low AMSTAR-2 rating typically used few pain-associated risks of bias, whereas those providing moderate or high confidence used more.

The implication is that significant sources of potential bias were likely to have affected results from most of the 57 systematic reviews. Cochrane reviews tended to be better, especially more recent reviews, whereas conflicted systematic reviews with pharmaceutical company backing were poor.

Quality was not associated with the impact of reviews as judged by annual citation rates on Google Scholar. Three of the top 5 cited reviews met many pain criteria (although not necessarily AMSTAR); the strength of recommendation in the abstracts of these reviews was weak negative or none. Two were rated critically low confidence by AMSTAR and used very few important criteria for avoiding bias in pain studies and reviews (Table 4); their strength of recommendation was moderately or strongly positive.

Of the 17 reviews attempting a numerical calculation of the magnitude of the analgesic effect, 9 had a positive and 5 a negative recommendation. Reviews with positive recommendations were associated with a statistically significant analgesic effect, but not reviews with negative or no recommendation. Susceptibility to publication bias for 3 reviews with statistical significance showed that the result would have been overturned by a null result from a clinical trial of 100 to 250 patients. Moreover, 8 of 19 reviews with moderate or strong positive assessment, whether or not they found a significant benefit, did not evaluate risk of bias, whereas all making a negative recommendation assessed risk of bias.

To summarise, what we have is a body of work that tells us little about whether any particular cannabinoid or cannabis-based treatment tested to date, at a particular dose and route of administration, given to someone with a particular form of pain could lead to a particular degree of pain reduction (at least 50% pain intensity reduction or reduction of pain to just mild54). Low-quality reviews do no more than suggest there may be, whereas the highest quality say probably not.

It is telling that a U.S. National Academies of Sciences, Engineering and Medicine report on therapeutic effects of cannabis and cannabinoids, and a later update,1,18 concluded that there is “substantial” evidence that cannabis is an effective treatment for chronic pain in adults. The committee included experts in substance abuse, cardiovascular health, epidemiology, immunology, pharmacology, pulmonary health, neurodevelopment, oncology, pediatrics, public health, and systematic review methodology, but not pain. It based much of its findings on pain on the systematic review of Whiting et al.84 That review was given an AMSTAR rating of critically low confidence and used only 2/8 pain methodologies. Moreover, for the patient-orientated outcome of at least 30% pain intensity reduction, it reported a result not significantly different from placebo, including, as it did, no significant difference in the 95% confidence interval (odds ratio 1.41 [0.99-2.00]). That conclusion should be revisited, revised, or retracted, because it is significantly misleading.

There are several lessons:

  • (1) The label of systematic review does not itself confer value for pain. Generic scoring systems for systematic reviews provide limited confidence, and the best mechanism to ensure that a systematic review provides a robust and reliable answer is to combine generic Cochrane approaches with pain-specific criteria, as several Cochrane review groups do. Systematic reviews of cannabinoids, cannabis, and CBM in pain require authors skilled not only in systematic review methodology but also those knowledgeable about pain and cannabinoids.
  • (2) Clinical trials to measure analgesic effect have a long-established basic methodology, but recent decades have demonstrated additional situation-specific factors needing consideration. Trials need to be conducted to the highest standards (especially those for registration or marketing purposes) and provide outcomes of clinical as well as statistical relevance to both efficacy and harms.
  • (3) RCTs investigating cannabis, cannabinoids, and CBM and pain should be designed to include well-defined populations with specific pain diagnoses, evaluate particular interventions (specific cannabinoid, doses, and route of administration) and comparators, and report on meaningful patient-reported pain outcomes (including functional outcomes and not just analgesic efficacy). Good RCTs can be complemented by patient registries to gather data on long-term patient outcomes to explore effectiveness of cannabinoids, cannabis, and CBMs for the treatment of pain and function in real life.
  • (4) More details of the clinical trials should be provided in reviews, particularly relating to concomitant analgesic medication, previous use of cannabinoids and other analgesics, and whether testing has been conducted to exclude nontrial use of drugs in test or placebo arms.
  • (5) The fact that on AMSTAR-2 alone, 41 systematic reviews providing critically low confidence in their results were published in medical journals (including some prestigious journals) indicates a potential problem with research quality. It is debatable whether this is a failure of journals and the peer review system, or whether scoring systems are unrealistic and penalise reviews unnecessary. Does a literature search of gray literature actually improve systematic reviews of RCTS, for example?
  • (6) Low quality and overclaiming positive benefits have long been associated.37,39 There is no obvious sign of improvement, and that is a matter of considerable concern.
  • (7) The link between the low quality of reviews and the positive or negative assessment of analgesic efficacy in review abstracts (often the only part that is read in any detail, or at all) is of concern. It begs the question not of challenge to these reviews, but to whether the implications are such as to consider calls for retraction. A Cochrane review compromised by methodological faults would likely be retracted, perhaps the only situation in which that would happen.

5. Conclusion

The primary reasons for this overview review were to examine the quality of the extant review literature and question whether a new systematic review would be needed. The results of the overview demonstrate that most reviews are lacking in quality and cannot provide a basis for decision-making. In the circumstances, a new systematic review adhering strictly to methodological requirements of AMSTAR and pain studies is required.

Conflict of interest statement

C. Eccleston reports grants from vs Arthritis, MayDay Foundation, Cochrane, and NIHR outside the submitted work. D.P. Finn reports grants from Alkermes Inc and Shionogi Ltd, outside the submitted work. N.B. Finnerup reports personal fees from Novartis Pharma, personal fees from Mitshubishi Tanabe Pharma, personal fees from Merck, personal fees from Almirall, personal fees from NeuroPN, and grants from EU PainCare, outside the submitted work. I. Gilron reports he is a Council Member of the International Association for the Study of Pain, as is part of the Presidential Task Force on Cannabis and Cannabinoid Analgesia, personal fees from Adynxx, personal fees from Biogen, personal fees from Eupraxia, personal fees from Novaremed, nonfinancial support from Canopy Health, nonfinancial support from Toronto Poly Clinic, and nonfinancial support from CannTrust, outside the submitted work. S. Haroutounian reports grants from Pfizer Inc and Disarm Therapeutics, and personal fees from Medoc Ltd and Rafa laboratories, outside the submitted work. A. Rice is a Council Member of IASP and Chair of the Presidential Task Force of the IASP. He undertook consultancy and advisory board work for Imperial College Consultants in the last 24 months; this has included personally remunerated work outside of the submitted work for: Pharmanovo, Lateral, Novartis, Pharmaleads, Mundipharma, Orion, Toray, Abide, Asahi Kasei, and Theranexis. He was the owner of share options in Spinifex Pharmaceuticals from which personal benefit accrued between 2015 and 2019 upon the acquisition of Spinifex by Novartis. Prof Rice is a named inventor on the patents—A.S.C. Rice, Vandevoorde S. and Lambert D. M Methods using N-(2propenyl)hexadecanamide and related amides to relive pain. WO2005/079771 pending, and Okuse et al. Methods of treating pain by inhibition of vgf activity EP13702262.0/WO2013110945 pending. During the conduct of the study, Imperial College received grants funding to support Prof Rice's programme of research from Biotechnology and Biological Sciences Research Council (BBSRC), Medical Research Council (MRC), Wellcome Trust, Alana and Sheila Diamond Charitable Trust, British Pain Society, Royal British Legion, and the European Commission (IMI2 (EQIPD); FP7 (Neuropain) and H2020 (Dolorisk)). M. Rowbotham reports personal fees from Adynxx, personal fees and other from CODA Biotherapeutics, and personal fees and other from SiteOne Therapeutics, outside the submitted work; and none of the entities listed are developing cannabinoid or cannabis-based medicines. M. Wallace Wallace reports personal fees from Insys, outside the submitted work. The remaining authors have no conflicts of interest to declare.

Appendix A. Supplemental digital content

Supplemental digital content associated with this article can be found online at https://links.lww.com/PAIN/B61.

Acknowledgements

This work is part of the effort of the IASP Presidential Taskforce on cannabis and cannabinoid analgesia. The authors thank Alexandra Fogarty for her help with searching.

The International Association for the Study of Pain commissioned this work in the form of a Presidential Task Force and funded attendance for the authors at working meeting in Washington DC, November 2019.

References

[1]. Abrams DI. The therapeutic effects of cannabis and cannabinoids: an update from the national Academies of sciences, engineering and medicine report. Eur J Intern Med 2018;49:7–11.
[2]. AlBalawi Z, McAlister FA, Thorlund K, Wong M, Wetterslev J. Random error in cardiovascular meta-analyses: how common are false positive and false negative results?. Int J Cardiol 2013;168:1102–7.
[3]. Alessa M, Witham D, Morriss-Roberts C. The effectiveness and safety of cannabis/cannabinoids for painful diabetic neuropathy: a systematic review. Glob Scientific Res J Diabetes 2018;1:9–20.
[4]. Almeida MO, Yamato TP, Parreira PDCS, Costa LOP, Kamper S, Saragiotto BT. Overall confidence in the results of systematic reviews on exercise therapy for chronic low back pain: a cross-sectional analysis using the Assessing the Methodological Quality of Systematic Reviews (AMSTAR) 2 tool. Braz J Phys Ther 2020;24:103–17.
[5]. Alviar MJM, Hale T, Lim-Dungca M. Pharmacologic interventions for treating phantom limb pain. Cochrane Database Syst Rev 2016;10:CD006380.
[6]. Amato L, Minozzi S, Mitrova Z, Parmelli E, Saulle R, Cruciani F, Vecchi S, Davoli M. Systematic review of safeness and therapeutic efficacy of cannabis in patients with multiple sclerosis, neuropathicpain, and in oncological patients treated with chemotherapy. Epidemiol Prev 2017;41:279–93.
[7]. Andreae MH, Carter GM, Shaparin N, Suslov K, Ellis RJ, Ware MA, Abrams DI, Prasad H, Wilsey B, Indyk D, Johnson M, Sacks HS. Inhaled cannabis for chronic neuropathic pain: a meta-analysis of individual patient data. J Pain 2015;16:1221–32.
[8]. Artukoglu BB, Beyer C, Zuloff-Shani A, Brener E, Bloch MH. Efficacy of palmitoylethanolamide for pain: a meta-analysis. Pain Physician 2017;20:353–62.
[9]. Aviram J, Samuelly-Leichtag G. Efficacy of cannabis-based medicines for pain management: a systematic review and meta-analysis of randomized controlled trials. Pain Physician 2017;20:E755–96.
[10]. Baldinger R, Katzberg HD, Weber M. Treatment for cramps in amyotrophic lateral sclerosis/motor neuron disease. Cochrane Database Syst Rev 2012;4:CD004157.
[11]. Basinski H, Jensen HB, Stenager E. Der er evidens for brug af cannabinoider til symptomatisk behandling af multipel sklerose. Ugeskr Læger 2014;176:V09130552.
[12]. Ben Amar M. Cannabinoids in medicine: a review of their therapeutic potential. J Ethnopharmacol 2006;105:1–25.
[13]. Boland EG, Bennett MI, Allgar V, Boland JW. Cannabinoids for adult cancer-related pain: systematic review and meta-analysis. BMJ Support Palliat Care 2020;10:14-24.
[14]. Boychuk DG, Goddard G, Mauro G, Orellana MF. The effectiveness of cannabinoids in the management of chronic nonmalignant neuropathic pain: a systematic review. J Oral Facial Pain Headache 2015;29:7–14.
[15]. Brettschneider J, Kurent J, Ludolph A. Drug therapy for pain in amyotrophic lateral sclerosis or motor neuron disease. Cochrane Database Syst Rev 2013;6:CD005226.
[16]. Brok J, Thorlund K, Wetterslev J, Gluud C. Apparently conclusive meta-analyses may be inconclusive-Trial sequential analysis adjustment of random error risk due to repetitive testing of accumulating data in apparently conclusive neonatal meta-analyses. Int J Epidemiol 2009;38:287–98.
[17]. Campbell FA, Tramer MR, Carroll D, Reynolds DJ, Moore RA, McQuay HJ. Are cannabinoids an effective and safe treatment option in the management of pain? A qualitative systematic review. BMJ 2001;323:13–6.
[18]. Committee on the Health Effects of Marijuana: An Evidence Review and Research Agenda; Board on Population Health and Public Health Practice; Health and Medicine Division; National Academies of Sciences, Engineering, and Medicine. The health effects of cannabis and cannabinoids: the current state of evidence and recommendations for research, 2017. Available at: http://nap.edu/24625. Accessed March 1, 2020.
[19]. da Rovare VP, Magalhães GPA, Jardini GDA, Beraldo ML, Gameiro MO, Agarwal A, Luvizutto GJ, Paula-Ramos L, Camargo SEA, de Oliveira LD, Bazan R, El Dib R. Cannabinoids for spasticity due to multiple sclerosis or paraplegia: a systematic review and meta-analysis of randomized clinical trials. Complement Ther Med 2017;34:170–85.
[20]. Darkovska-Serafimovska M, Serafimovska T, Arsova-Sarafinovska Z, Stefanoski S, Keskovski Z, Balkanov T. Pharmacotherapeutic considerations for use of cannabinoids to relieve pain in patients with malignant diseases. J Pain Res 2018;11:837–42.
[21]. de Souza Nascimento S, Desantana JM, Nampo FK, Ribeiro EA, da Silva DL, Araújo-Júnior JX, da Silva Almeida JR, Bonjardim LR, de Souza Araújo AA, Quintans-Júnior LJ. Efficacy and safety of medicinal plants or related natural products for fibromyalgia: a systematic review. Evid Based Complement Alternat Med 2013;2013:149468.
[22]. Dechartres A, Altman DG, Trinquart L, Boutron I, Ravaud P. Association between analytic strategy and estimates of treatment outcomes in meta-analyses. Jama 2014;312:623–30.
[23]. Dechartres A, Trinquart L, Boutron I, Ravaud P. Influence of trial sample size on treatment effect estimates: meta-epidemiological study. Bmj 2013;346:f2304.
[24]. Deshpande A, Mailis-Gagnon A, Zoheiry N, Lakha SF. Efficacy and adverse effects of medical marijuana for chronic noncancer pain: systematic review of randomized controlled trials. Can Fam Physician 2015;61:e372–81.
[25]. EPOC Anon. EPOC author resources. Available at: epoc.cochrane.org/epoc-specific-resources-review-authors 2015. Accessed January 20, 2018.
[26]. Fanelli D, Costas R, Ioannidis JP. Meta-assessment of bias in science. Proc Natl Acad Sci U S A 2017;114:3714–19.
[27]. Finnerup NB, Attal N, Haroutounian S, McNicol E, Baron R, Dworkin RH, Gilron I, Haanpää M, Hansson P, Jensen TS, Kamerman PR, Lund K, Moore A, Raja SN, Rice AS, Rowbotham M, Sena E, Siddall P, Smith BH, Wallace M. Pharmacotherapy for neuropathic pain in adults: a systematic review and meta-analysis. Lancet Neurol 2015;14:162–73.
[28]. Fisher E, Eccleston C, Degenhardt L, Finn DP, Finnerup NB, Gilron I, Haroutounian S, Krane E, Rice ASC, Rowbotham M, Wallace M, Moore RA. Cannabinoids, cannabis, and cannabis-based medicine for pain management: a protocol for an overview of systematic reviews and a systematic review of randomised controlled trials. Pain Rep 2019;4:e741.
[29]. Fitzcharles MA, Ste-Marie PA, Häuser W, Clauw DJ, Jamal S, Karsh J, Landry T, Leclercq S, Mcdougall JJ, Shir Y, Shojania K, Walsh Z. Efficacy, tolerability, and safety of cannabinoid treatments in the rheumatic diseases: a systematic review of randomized controlled trials. Arthritis Care Res (Hoboken) 2016;68:681–8.
[30]. Fleming PS, Koletsi D, Ioannidis JP, Pandis N. High quality of the evidence for medical and other health-related interventions was uncommon in Cochrane systematic reviews. J Clin Epidemiol 2016;78:34–42.
[31]. Gilron I, Blyth FM, Degenhardt L, Di Forti M, Eccleston C, Haroutounian S, Moore A, Rice ASC, Wallace M. Risks of harm with cannabinoids, cannabis, and cannabis-based medicine for pain management relevant to patients receiving pain treatment: protocol for an overview of systematic reviews. Pain Rep 2019;4:e742.
[32]. Harrison AM, Heritier F, Childs BG, Bostwick JM, Dziadzko MA. Systematic review of the use of phytochemicals for management of pain in cancer therapy. Biomed Res Int 2015;2015:506327.
[33]. Häuser W, Welsch P, Klose P, Radbruch L, Fitzcharles MA. Efficacy, tolerability and safety of cannabis-based medicines for cancer pain: a systematic review with meta-analysis of randomised controlled trials. Schmerz 2019;33:424–36.
[34]. Higgins JPT, Green S. cochrane handbook for systematic reviews of interventions, 2011. Available at: https://handbook-5-1.cochrane.org/front_page.htm. Accessed February 15, 2020.
[35]. Hou S, Huh B, Kim HK, Kim KH, Abdi S. Treatment of chemotherapy-induced peripheral neuropathy: systematic review and recommendations. Pain Physician 2018;21:571–92.
[36]. IntHout J, Ioannidis JP, Borm GF, Goeman JJ. Small studies are more heterogeneous than large ones: a meta-meta-analysis. J Clin Epidemiol 2015;68:860–9.
[37]. Ioannidis JP. Why most published research findings are false. Plos Med 2005;2:e124.
[38]. Iskedjian M, Bereza B, Gordon A, Piwko C, Einarson TR. Meta-analysis of cannabis based treatments for neuropathic and multiple sclerosis-related pain. Curr Med Res Opin 2007;23:17–24.
[39]. Jadad AR, McQuay HJ. Meta-analyses to evaluate analgesic interventions: a systematic qualitative review of their methodology. J Clin Epidemiol 1996;49:235–43.
[40]. Jawahar R, Oh U, Yang S, Lapane KL. A systematic review of pharmacological pain management in multiple sclerosis. Drugs 2013;73:1711–22.
[41]. Jørgensen AW, Hilden J, Gøtzsche PC. Cochrane reviews compared with industry supported meta-analyses and other meta-analyses of the same drugs: systematic review. BMJ 2006;333:782.
[42]. Koppel BS, Brust JC, Fife T, Bronstein J, Youssof S, Gronseth G, Gloss D. Systematic review: efficacy and safety of medical marijuana in selected neurologic disorders: report of the Guideline Development Subcommittee of the American Academy of Neurology. Neurology 2014;82:1556–63.
[43]. Lynch ME, Campbell F. Cannabinoids for treatment of chronic non-cancer pain; a systematic review of randomized trials. Br J Clin Pharmacol 2011;72:735–44.
[44]. Lynch ME, Ware MA. Cannabinoids for the treatment of chronic non-cancer pain: an updated systematic review of randomized controlled trials. J Neuroimmune Pharmacol 2015;10:293–301.
[45]. Macfarlane GJ, El-Metwally A, De Silva V, Ernst E, Dowds GL, Moots RJ. Arthritis Research UK Working Group on Complementary and Alternative Medicines. Evidence for the efficacy of complementary and alternative medicines in the management of rheumatoid arthritis: a systematic review. Rheumatology (Oxford) 2011;50:1672–83.
[46]. Martín-Sánchez E, Furukawa TA, Taylor J, Martin JL. Systematic review and meta-analysis of cannabis treatment for chronic pain. Pain Med 2009;10:1353–68.
[47]. Mehta S, McIntyre A, Janzen S, Loh E, Teasell R. Spinal cord injury rehabilitation evidence team. Systematic review of pharmacologic treatments of pain after spinal cord injury: an update. Arch Phys Med Rehabil 2016;97:1381–91.e1.
[48]. Meng H, Johnston B, Englesakis M, Moulin DE, Bhatia A. Selective cannabinoids for chronic neuropathic pain: a systematic review and meta-analysis. Anesth Analg 2017;125:1638–52.
[49]. Moher D, Shamseer L, Clarke M, Ghersi D, Liberati A, Petticrew M, Shekelle P, Stewart LA, Group PP. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev 2015;4:1.
[50]. Mohiuddin MM, Mizubuti GB, Haroutounian S, Smith SM, Rice ASC, Campbell F, Park R, Gilron I. Adherence to consolidated standards of reporting trials (CONSORT) guidelines for reporting safety outcomes in trials of medical cannabis and cannabis-based medicines for chronic noncancer pain: a systematic review. Clin J Pain 2020 36:302–19.
[51]. Moore RA, Barden J, Derry S, McQuay HJ. Managing potential publication bias. In: McQuay HJ, Kalso E, Moore RA, editors. Systematic reviews in pain research: methodology refined seattle: IASP Press, 2008. pp. 15–23.
[52]. Moore RA, Gavaghan D, Tramèr MR, Collins SL, McQuay HJ. Size is everything--large amounts of information are needed to overcome random effects in estimating direction and magnitude of treatment effects. PAIN 1998;78:209–16.
[53]. Moore RA, Straube S, Aldington D. Pain measures and cut-offs—“no worse than mild pain” as a simple, universal outcome. Anaesthesia 2013;68:400–12.
[54]. Mücke M, Phillips T, Radbruch L, Petzke F, Häuser W. Cannabis-based medicines for chronic neuropathic pain in adults. Cochrane Database Syst Rev 2018;3:CD012182.
[55]. Mücke M, Weier M, Carter C, Copeland J, Degenhardt L, Cuhls H, Radbruch L, Häuser W, Conrad R. Systematic review and meta-analysis of cannabinoids in palliative medicine. J Cachexia Sarcopenia Muscle 2018;9:220–34.
[56]. NICE guideline. Cannabis-based medicinal products. 2019 Available at: www.nice.org.uk/guidance/ng144. Accessed March 1, 2020.
[57]. Nüesch E, Trelle S, Reichenbach S, Rutjes AW, Tschannen B, Altman DG, Egger M, Jüni P. Small study effects in meta-analyses of osteoarthritis trials: meta-epidemiological study. BMJ 2010;341:c3515.
[58]. Nugent SM, Morasco BJ, O'Neil ME, Freeman M, Low A, Kondo K, Elven C, Zakher B, Motu'apuaka M, Paynter R, Kansagara D. The effects of cannabis among adults with chronic pain and an overview of general harms: a systematic review. Ann Intern Med 2017;167:319–31.
[59]. Oltean H, Robbins C, van Tulder MW, Berman BM, Bombardier C, Gagnier JJ. Herbal medicine for low-back pain. Cochrane Database Syst Rev 2014;12:CD004504.
[60]. Paladini A, Fusco M, Cenacchi T, Schievano C, Piroli A, Varrassi G. Palmitoylethanolamide, a special food for medical purposes, in the treatment of chronic pain: a pooled data meta-analysis. Pain Physician 2016;19:11–24.
[61]. Phillips TJ, Cherry CL, Cox S, Marshall SJ, Rice AS. Pharmacological treatment of painful HIV-associated sensory neuropathy: a systematic review and meta-analysis of randomised controlled trials. PLoS One 2010;5:e14433.
[62]. Pittler MH, Ernst E. Complementary therapies for neuropathic and neuralgic pain: systematic review. Clin J Pain 2008;24:731–3.
[63]. Qureshi AR, Rana AQ, Malik SH, Rizvi SFH, Akhter S, Vannabouathong C, Sarfraz Z, Rana R. Comprehensive examination of therapies for pain in Parkinson's disease: a systematic review and meta-analysis. Neuroepidemiology 2018;51:190–206.
[64]. Richards BL, Whittle SL, Buchbinder R. Neuromodulators for painmanagement in rheumatoid arthritis. Cochrane Database Syst Rev 2012;1:CD008921.
[65]. Roberts I, Ker K, Edwards P, Beecher D, Manno D, Sydenham E. The knowledge system underpinning healthcare is not fit for purpose and must change. BMJ 2015;350:h2463.
[66]. Seers T, Derry S, Seers K, Moore RA. Professionals underestimate patients' pain: a comprehensive review. PAIN 2018;159:811–18.
[67]. Shea BJ, Reeves BC, Wells G, Thuku M, Hamel C, Moran J, Moher D, Tugwell P, Welch V, Kristjansson E, Henry DA. Amstar 2: a critical appraisal tool for systematic reviews that include randomised or non-randomised studies of healthcare interventions, or both. BMJ 2017;358:j4008.
[68]. Smith R. Where is the wisdom…? BMJ 1991;303:798–9.
[69]. Snedecor SJ, Sudharshan L, Cappelleri JC, Sadosky A, Desai P, Jalundhwala Y, Botteman M. Systematic review and meta-analysis of pharmacological therapies for pain associated with postherpetic neuralgia and less common neuropathic conditions. Int J Clin Pract 2014;68:900–18.
[70]. Snedecor SJ, Sudharshan L, Cappelleri JC, Sadosky A, Desai P, Jalundhwala YJ, Botteman M. Systematic review and comparison of pharmacologic therapies for neuropathic pain associated with spinal cord injury. J Pain Res 2013;6:539–47.
[71]. Snedecor SJ, Sudharshan L, Cappelleri JC, Sadosky A, Mehta S, Botteman M. Systematic review and meta-analysis of pharmacological therapies for painful diabetic peripheral neuropathy. Pain Pract 2014;14:167–84.
[72]. Soliman N, Hohmann AG, Haroutounian S, Wever K, Rice ASC, Finn DP. A protocol for the systematic review and meta-analysis of studies in which cannabinoids were tested for antinociceptive effects in animal models of pathological or injury-related persistent pain. PAIN Rep 2019;4:e766.
[73]. Stevens AJ, Higgins MD. A systematic review of the analgesic efficacy of cannabinoid medications in the management of acute pain. Acta Anaesthesiol Scand 2017;61:268–80.
[74]. Stockings E, Campbell G, Hall WD, Nielsen S, Zagic D, Rahman R, Murnion B, Farrell M, Weier M, Degenhardt L. Cannabis and cannabinoids for the treatment of people with chronic noncancer pain conditions: a systematic review and meta-analysis of controlled and observational studies. PAIN 2018;159:1932–54.
[75]. Tateo S. State of the evidence: cannabinoids and cancer pain-A systematic review. J Am Assoc Nurse Pract 2017;29:94–103.
[76]. Thorlund K, Imberger G, Walsh M, Chu R, Gluud C, Wetterslev J, Guyatt G, Devereaux PJ, Thabane L. The number of patients and events required to limit the risk of overestimation of intervention effects in meta-analysis--a simulation study. PLoS One 2011;6:e25491.
[77]. Torres-Moreno MC, Papaseit E, Torrens M, Farré M. Assessment of efficacy and tolerability of medicinal cannabinoids in patients with multiple sclerosis: a systematic review and meta-analysis. JAMA Netw Open 2018;1:e183485.
[78]. Tsang CC, Giudice MG. Nabilone for the management of pain. Pharmacotherapy 2016;36:273–86.
[79]. Turner RM, Bird SM, Higgins JP. The impact of study size on meta-analyses: examination of underpowered studies in Cochrane reviews. PLoS One 2013;8:e59202.
[80]. van den Beuken-van Everdingen MH, de Graeff A, Jongen JL, Dijkstra D, Mostovaya I, Vissers KC. National guideline working group “Diagnosis treatment of cancer pain”. Pharmacological treatment of pain in cancer patients: the role of adjuvant analgesics, a systematic review. Pain Pract 2017;17:409–19.
[81]. Volz MS, Siegmund B, Häuser W. Efficacy, tolerability, and safety of cannabinoids in gastroenterology. A systematic review [Wirksamkeit, Vertr.glichkeit und Sicherheit von Cannabinoiden in der Gastroenterologie: eine systematische bersichtsarbeit.]. Schmerz 2016;30:37–46.
[82]. Walitt B, Klose P, Fitzcharles MA, Phillips T, Häuser W. Cannabinoids for fibromyalgia. Cochrane Database Syst Rev 2016;7:CD011694.
[83]. Watson CP, Gilron I, Sawynok J. A qualitative systematic review of head-to-head randomized controlled trials of oral analgesics in neuropathic pain. Pain Res Manag 2010;15:147–57.
[84]. Whiting PF, Wolff RF, Deshpande S, Di Nisio M, Duffy S, Hernandez AV, Keurentjes JC, Lang S, Misso K, Ryder S, Schmidlkofer S, Westwood M, Kleijnen J. Cannabinoids for medical use: a systematic review and meta-analysis. JAMA 2015;313:2456–73.
[85]. Wong SS, Wilens TE. Medical cannabinoids in children and adolescents: a systematic review. Pediatrics 2017;140:e20171818.
[86]. Yanes JA, McKinnell ZE, Reid MA, Busler JN, Michel JS, Pangelinan MM, Sutherland MT, Younger JW, Gonzalez R, Robinson JL. Effects of cannabinoid administration for pain: a meta-analysis and meta-regression. Exp Clin Psychopharmacol 2019;27:370–82.
[87]. Zhornitsky S, Potvin S. Cannabidiol in humans-the quest for therapeutic targets. Pharmaceuticals (Basel) 2012;5:529–52.

Supplemental Digital Content

© 2020 International Association for the Study of Pain