Chronic pain is associated with enormous personal, social, and economic burden and is the largest contributor to years lived with disability globally (Rice et al, 2015). Despite this, existing medications provide only modest relief. Opioids in particular have considerable side effects, including constipation, impaired sleep, and respiratory depression (Chou et al, 2015). The last two decades have seen an increase in the prescription of opioids, which has been associated with an increase in opioid use disorders and opioid-related mortality (Chou et al, 2015; Volkow and McLellan, 2016; Zedler et al, 2014). This has been termed as an ‘opioid crisis’, and has caused regulators, health professionals, and the public to begin seeking means to reduce problems associated with high-dose opioid use. Consequently, there is a need for evidence-based strategies for reducing reliance on high-dose opioids without compromising pain management.

Using combinations of medications to harness complementary but distinct mechanisms of action can maximize the analgesic response, enabling the use of a lower dose of each medication and resulting in an improved side effect profile. One promising area for medication combinations is the use of opioid-sparing medications. Opioid-sparing medications, when co-administered with opioids, enable a reduced opioid dose without loss of analgesic efficacy. Cannabinoid medications are increasingly being studied for their analgesic- and opioid-sparing potential. The endocannabinoid system represents an ideal target because it is a key endogenous system in modulating pain-processing pathways (Woodhams et al, 2015).

The endocannabinoid system is composed of the cannabinoid CB1 and CB2 receptors, the endocannabinoid ligands anandamide and 2-arachidonoylglycerol, and their synthesis and degradation system (Pertwee, 2006). CB1 and CB2 receptors are differentially expressed on the central nervous system (Cencioni et al, 2010; Herkenham et al, 1991) and play important roles in pain processes. Both cannabinoid receptors and endocannabinoids are present in the primary afferent pain circuits to the brain (Manzanares et al, 1999; Woodhams et al, 2015). Cannabinoid and opioid receptors have similar signal transduction systems (Cichewicz, 2004; Howlett et al, 2002; Vigano et al, 2005) and are expressed in several brain regions involved in antinociception, including the periaqueductal gray, raphe nuclei, and central-medial thalamic nuclei (Cichewicz, 2004). In addition, mu-opioid receptors and CB1 receptors co-localize in the spinal cord at the first synaptic contact for peripheral nociceptive afferent neurons (Hohmann et al, 1999; Salio et al, 2001).

It has previously been observed that CB2 receptors indirectly stimulate opioid receptors located in primary afferent pathways (Ibrahim et al, 2005). Therefore, in addition to their direct analgesic effects, cannabinoids may work synergistically to enhance opioid analgesia. The behavioral, anatomical, and biochemical similarities between opioid and cannabinoid receptor systems and their endogenous ligands are well documented. For example, activation of either cannabinoid or opioid receptors produces comparable neurobehavioral and physiological effects, including antinociception (Manzanares et al, 1999). This is highlighted by both CB1 and CB2 agonists being able to induce antinociception by increasing opioid precursors’ gene expression or via release of endogenous opioids (Houser et al, 2000; Ibrahim et al, 2005; Valverde et al, 2001). Further, pharmacological modulation of the opioid system can modify the effects of delta-9-tetrahydrocannabinol (delta-9-THC)—a partial agonist at the CB1 and CB2 receptor—on nociception (Mason et al, 1999; Pugh et al, 1997; Smith et al, 1994) and vice versa. Finally, cannabinoid antagonists have been shown to reverse the antinociception induced by morphine (da Fonseca Pacheco et al, 2008). Collectively, this strongly supports shared mechanisms between both systems in regard to analgesia.

Animal models have identified a role for CB1 receptor activation in reducing neuropathic, visceral, and inflammatory pain (Pertwee, 2008; Walker et al, 1999). Several pre-clinical studies have demonstrated that systemic administration of cannabinoid receptor ligands produces analgesia in acute and chronic pain models (Walker and Huang, 2002). In addition, the role of CB2 receptors has been explored in pre-clinical studies, suggesting that these receptors may mediate effects in inflammatory pain states (Ibrahim et al, 2006; Quartilho et al, 2003), and reduce inflammation and neuropathic pain (Gui et al, 2015).

Further to these pre-clinical findings, clinical studies indicate that cannabinoid administration may reduce pain and improve other symptoms such as sleep disturbances associated with chronic pain (Ware et al, 2010a; Ware et al, 2010b). This effect could be mediated by delta-9-THC, which is the main psychoactive ingredient present in cannabis (Cichewicz, 2004; Jensen et al, 2015). Despite the growing body of relevant literature, to date no systematic review has focused on the opioid-sparing effects of cannabinoids. To address this gap, we conducted a systematic review of pre-clinical and clinical studies to examine the strength of existing evidence demonstrating the opioid-sparing effect of cannabinoids in the context of analgesia.

Materials and methods


We conducted a systematic search of the literature in accordance with recommendations by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) (Moher et al, 2009). The search aimed to identify clinical and pre-clinical studies using the following electronic databases: Scopus, Cochrane Database of Systematic Reviews, Medline, and Embase. Search terms are listed and a sample search strategy is reported in Supplementary Appendix 1. No date limits were included. Searches were run on 29 October 2015. In addition, reference lists from identified studies and review articles were searched to find additional studies not identified by the main search.

Eligible studies included:

  • Human or animal studies.

  • Outcomes of either pain/analgesia or opioid requirements/opioid-sparing effects from concurrently administered opioids and cannabinoids.

  • Controlled clinical studies and case series.

Titles were screened by two authors (SN and PS). Where inconsistencies were identified, the authors were able to reach consensus on each occasion.

Data Extraction and Outcomes

Data extraction forms were developed and circulated to the author group before piloting and refining. All data were extracted by one of the authors (SN, PS, or JMT) and checked by a second author (SN, PS, or JMT). These same authors reviewed and resolved any inconsistencies, with input from the authorship group as required. When required data were missing, attempts were made to contact authors of published reports to collect additional information.

Outcome Measures

For pre-clinical studies, the primary outcome was the dose of opioid required to give an equivalent antinociceptive effect in the presence and absence of cannabinoids. For clinical studies, the primary outcome was evidence of the opioid-sparing effect of cannabinoids. Data were extracted on opioid dose and/or analgesic outcome where cannabinoids were co-administered. Secondary outcome measures examined included analgesia, sleep, and quality of life.


Pre-clinical studies

Data were extracted and a narrative review was conducted. Ten studies were identified as sufficiently similar in design and outcome measures to be eligible for meta-analysis. Of these, six reported sufficient data to enable meta-analysis; that is, the dose of opioid required to produce comparable analgesia in the presence and absence of cannabinoids, the variance of the observed dose, and the sample size. Authors of the other studies were contacted in an attempt to include additional studies in the meta-analysis; however, no additional data were identified to enable the inclusion of any additional studies.

To prepare the data for the meta-analysis, the effective dose (ED50) and either confidence limits or SE were extracted from the relevant literature. ED50 is calculated on the log10 scale. Therefore, to meet the assumption of normality, the log10 must be used in the meta-analysis. The log10 of the confidence limits must also be determined to calculate the SD of the log10:

where UL is the upper confidence limit.

When only SE was reported, the confidence limits were calculated using the method of Litchfield and Wilcoxon (1949) and the above procedure was repeated to calculate the SD. This method also allowed for the inclusion of studies that did not report exact sample sizes for all treatment groups, as sample size was not required for the calculation of SD.

Data for the meta-analysis were analyzed using Review Manager 5.1 (Cochrane Collaboration, Oxford, UK). When calculating the continuous outcome of an equally effective opioid dose (eg, the log10ED50 for morphine when administered alone vs when administered with a cannabinoid), the inverse variance statistical method and random effects model were used to compensate for study heterogeneity.

No statistical difference was found in outcomes between the studies that used different species or nociceptive assays. Therefore, the mean difference of log 10ED50 of and the corresponding 95% confidence intervals (CI) were calculated. Due to the nature of log calculations, the mean difference—when back-transformed to the original units—represents the response ratio. For easier interpretation, we present the reciprocal of the response rate.

Assessment of bias in pre-clinical studies

A funnel plot was produced to examine publication bias and small study effects in the pre-clinical studies included in the meta-analysis.

Clinical studies

The nine clinical studies identified were heterogeneous in design and outcomes, and therefore not suitable for meta-analysis. Thus, a narrative synthesis was conducted instead, with all studies scored for quality using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) criteria (Guyatt et al, 2008).


The initial searches identified 3019 records after duplicates were removed, with 19 pre-clinical and nine clinical studies identified for inclusion in the final review (see Figure 1 for the PRISMA diagram).

Figure 1
figure 1

PRISMA diagram showing study identification. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

PowerPoint slide

Summary of Pre-Clinical Studies

Nineteen pre-clinical studies were identified in which the analgesic effect of opioid and cannabinoid co-administration was examined (Cichewicz et al, 1999, 2005; Cichewicz and McCarthy, 2003; Cox et al, 2007; Finn et al, 2004; Katsuyama et al, 2013; Li et al, 2008; Maguire et al, 2013; Pugh Jr et al, 1996; Reche et al, 1996; Smith et al, 1998, 2007; Tham et al, 2005; Wakley and Craft, 2011; Welch and Stevens, 1992; Williams et al, 2006, 2008; Wilson et al, 2008; Yesilyurt et al, 2003) (Table 1). Fourteen of these studies examined delta-9-THC, whereas one to two studies examined each of 10 other cannabinoid agonists, including beta-caryophyllene, CP 55 940, CP 56 667, delta-8-THC, 11-hydroxy-delta-9-THC, dextronantradol, levonantradol, WIN 55, 212-2, and HU-210. Seventeen studies examined morphine, three studies examined codeine, and one to two studies examined buprenorphine, fentanyl, oxycodone, morphine, hydromorphone, methadone, LAAM, meperidine, and pentazocine. Most of the studies used rodents; however, two used rhesus monkeys and one used guinea pigs. The most common antinociceptive assays were tail-flick tests (n=10) and hot plate tests (n=5), although individual studies also used other forms of mechanical, thermal, and chemical nociception.

Table 1 Summary of Evidence of Opioid-Sparing Effects from Pre-Clinical Studies

Most studies (17 of the 19) demonstrated that combining a cannabinoid with an opioid resulted in a synergistic effect on analgesia compared to the analgesic effects of the individual drugs. One study examined a single dose of morphine and demonstrated that morphine could potentiate the analgesic effect of intrathecally administered delta-9-THC (Reche et al, 1996). However, this study could not demonstrate an opioid-sparing effect due to the use of a single dose of opioid. Another study found that 2 mg/kg morphine administered with 1 mg/kg delta-9-THC resulted in a significant effect on nociception compared to morphine alone (p<005), but not when compared to delta-9-THC alone (Finn et al, 2004). In another study, a greater increase in hot plate latency was found for morphine combined with HU-210 (38.9 s±1.1 s) compared with HU-210 alone (33.1 s±4.0 s) (Wilson et al, 2008); however, this difference did not reach significance.

One study testing multiple opioid agonists identified clear synergistic effects for delta-9-THC for most opioid drugs, with the exception of fentanyl and pentazocine (Cichewicz et al, 1999). The potency ratio when administered alone for those opioids found to have a synergistic effect, compared to when those same opioids were co-administered with delta-9-THC varied between 2.2 and 25.8. Another study tested multiple cannabinoid agonists when co-administered with morphine and demonstrated a synergistic effect with delta-9-THC, delta-8-THC, levonantradol, and 11-hydroxy-delta-9-THC; additive effects with CP 55 940 and CP 56 667; and no observable potentiation of morphine effects with dextronantradol, which is an isomer of levonantradol (Welch and Stevens, 1992). In contrast to the finding of an additive effect for CP 55 940, two other studies of CP 55 940 in combinations with morphine demonstrated a synergistic analgesic effect (Maguire et al, 2013; Tham et al, 2005). In addition to changes in the magnitude of the analgesic effect, two studies showed that the duration of the analgesic effect can be extended by administrating a low-dose opioid and cannabinoid in combination, compared with administrating an opioid alone (Williams et al, 2006; Yesilyurt et al, 2003).

Meta-Analysis of Pre-Clinical Studies

Six studies used sufficiently similar approaches to enable a meta-analysis (Cichewicz et al, 1999; Cichewicz and Welch, 2003; Cox et al, 2007; Smith et al, 1998; Welch and Stevens, 1992; Williams et al, 2008) (Figure 2). A further four studies were comparable in study design, but did not contain the required data (ED50 or variance on estimates) to enable meta-analysis (Finn et al, 2004; Pugh Jr et al, 1996; Smith et al, 2007; Williams et al, 2006). All studies included in the meta-analysis used rodents and reported comparable antinociceptive doses of morphine alone and morphine co-administered with delta-9-THC. Results from the meta-analysis are reported in terms of mean difference.

Figure 2
figure 2

Forrest plot for meta-analysis examining the opioid-sparing effect of delta-9-THC when co-administered with morphine. Note: all mean difference and SD values are of log10ED50. THC, tetrahydrocannabinol.

PowerPoint slide

The meta-analysis identified a significant opioid-sparing effect with morphine and delta-9-THC co-administration, Z=5.59, p<0.001 (MD in log10ED50 –0.56 (–0.83, –0.29)). As there was significant heterogeneity in the data (I2=95%), a random effects model was used. When back-transformed to the original units, the response ratio was 3.6 (95% CI 1.95, 6.76), indicating that the median ED50 of morphine was 3.6 times lower when given in combination with delta-9-THC compared to when morphine was administered alone.

Two studies compared doses of codeine with and without delta-9-THC in rodents (Cichewicz et al, 1999; Cichewicz and Welch, 2003) (Figure 3). Both studies used male ICR mice and the tail-flick assay. Meta-analysis of these data indicated a significant opioid-sparing effect of delta-9-THC when co-administered with codeine, Z=2.49, p=0.01 (MD in the log10ED50 –0.98 (–1.76, –0.21)). Significant heterogeneity in the data (I2=98%) necessitated the use of a random effects model. When back-transformed to the original units, the response ratio was 9.5 (95% CI 1.6, 57.5), indicating that the ED50 of codeine was 9.5 times lower when given in combination with delta-9-THC compared to when codeine was administered alone.

Figure 3
figure 3

Forrest plot for meta-analysis examining the opioid-sparing effect of delta-9-THC when co-administered with codeine. Note: all mean difference and SD values are of log10ED50. THC, tetrahydrocannabinol.

PowerPoint slide

Funnel plots did not provide evidence of publication or small study bias with these pre-clinical studies (Figure 4).

Figure 4
figure 4

Funnel plot showing data from the six studies included in the meta-analysis. MD, mean difference, SE, standard error.

PowerPoint slide

Results from Clinical Studies

Nine clinical studies with 750 participants provided data relevant to the research question (Table 2); however, the heterogeneous nature of the studies precluded meta-analysis. Three laboratory-based studies examined pain responses in participants concurrently being administered opioids and cannabinoids. One study recruited people with mixed chronic non-cancer pain (n=24) who were prescribed opioids (Abrams et al, 2011). A significant reduction in pain ratings was observed for the participants in this study following co-administration of cannabinoids—39.6 (95% CI 35.8, 43.3) at baseline vs 29.1 (95% CI 25.4, 32.8) following co-administration (Abrams et al, 2011). It should be noted that no placebo or control condition was used in this study for comparison (Abrams et al, 2011).

Table 2 Summary of Evidence of Opioid-sparing Effects from Clinical Studies

In another two studies, healthy volunteers (n=12 and 13, respectively) participated in crossover studies, with single doses of placebo, morphine alone, dronabinol alone, and dronabinol and morphine combined administered over four sessions (Naef et al, 2003; Roberts et al, 2006). These studies did not identify a synergistic effect on experimental pain in healthy controls, although Roberts et al (2006) found that the co-administration of dronabinol and morphine resulted in a reduced unpleasantness of pain compared to either drug alone. In a case series examining the effects of cannabinoid administration in patients with chronic non-cancer pain, three patients with mixed pain conditions (multiple sclerosis, HIV-related peripheral neuropathy, and lower back and leg pain) reported reductions in opioid requirements after initiation of smoked cannabis plant material (Lynch and Clark, 2003).

Five controlled studies were identified. One small, non-randomized study of patients with advanced cancer pain found that 5 out of 12 patients achieved pain control after receiving a cannabis infusion, compared with 2 out of 14 achieving pain control in the control group—a non-statistically different effect (Lissoni et al, 2014). Two randomized controlled trials examined delta-9-THC : Cannabidiol (THC : CBD) combination oral sprays compared to a placebo (Johnson et al, 2010; Portenoy et al, 2012) in patients with cancer pain who were taking opioids. These studies found improved analgesia with the THC : CBD combination compared to the placebo. Johnson et al (2010) found no effect of THC : CBD on breakthrough opioid dose requirements. Portenoy et al (2012) conducted a dose-ranging study, using fixed dose ranges of the THC : CBD combination. In this study, a significant analgesic effect was only found in the lowest dose group, with poorer tolerability observed for higher doses.

Two controlled studies examined the effects of dronabinol: one in patients with mixed chronic pain (Narang et al, 2008) and one in patients with prostate cancer (Seeling et al, 2006). Narang et al (2008) found significantly reduced pain intensity with the opioid–cannabinoid combination in double-blinded laboratory sessions compared to opioid alone. Additional improvements in sleep, energy, and social functioning were reported in a 4-week open-label phase of the same study (Narang et al, 2008). In the study by Seeling et al (2006), perioperative use of dronabinol compared with a placebo in patients with prostate cancer, no difference was found in self-administered opioid dose requirements between groups.

Quality Ratings of Clinical Studies

The clinical studies were rated using the GRADE criteria. One study provided very-low-quality evidence, three studies provided low-quality evidence, two studies provided moderate-quality evidence, and three randomized controlled trials provided high-quality evidence. None of the high-quality studies provided evidence of an opioid-sparing effect. The only study that provided direct evidence of an opioid-sparing effect was rated as providing very low-quality evidence (Lynch and Clark, 2003).


Twenty-eight studies provided data relating to the potential opioid-sparing effect of cannabinoids in the context of opioid analgesia. Most of the pre-clinical studies examined reported reduced opioid requirements when co-administered with cannabinoids. Few controlled clinical studies measured opioid-sparing as an end point and findings relating to analgesia were mixed. Two controlled studies found no effect of cannabinoids on opioid dose requirements (Johnson et al, 2010; Seeling et al, 2006). One case series provided very low-quality evidence of a reduction in opioid dose requirements with cannabinoid co-administration (Lynch and Clark, 2003).

Most of the pre-clinical studies examined found synergistic effects when opioids and cannabinoids were co-administered, although two studies found that with specific opioids and cannabinoids the analgesic effect was additive rather than synergistic. Through meta-analyses, it was found that the doses of morphine and codeine required to produce the same analgesic effect were 3.6 and 9.5 times lower, respectively, when co-administered with delta-9-THC. Reductions in opioid requirements that are smaller than those seen in these pre-clinical studies may have relevance to pain treatment. Some confidence in these findings comes from the consistent observation of an opioid-sparing effect when using different nociceptive assays and in pain models of arthritis and diabetic neuropathy.

The relevance of the findings from these pre-clinical studies (with acute-dosing paradigms) to clinical chronic pain treatment must be considered. There are important limitations in translating findings from pre-clinical studies to clinical practice, particularly when evaluating doses and effect sizes. Although the outcomes of pre-clinical studies are often consistent with clinical studies, pre-clinical studies may over-represent effects. The lesser effect sizes in human studies have been attributed to the heterogeneity of clinical populations or the response being limited to sub-populations, reducing the overall effect observed (Berge, 2011). This underscores the importance of clinical studies to examine the effects found in pre-clinical work.

Controlled clinical studies demonstrated some beneficial effects of opioid and cannabinoid co-administration on outcomes of pain, sleep, and functioning in chronic pain patients (Abrams et al, 2011; Narang et al, 2008). One case series (n=3) provided very low-quality evidence of a reduction in opioid requirements with delta-9-THC administration. No randomized controlled studies were identified that provided evidence of an opioid-sparing effect of cannabinoids. Important limitations identified in these clinical studies included a lack of placebo control (Abrams et al, 2011; Lynch and Clark, 2003; Narang et al, 2008), difficulties extrapolating from experimental to clinical pain (Naef et al, 2003; Roberts et al, 2006), use of single doses (Naef et al, 2003; Roberts et al, 2006), use of small sample sizes (Lissoni et al, 2014; Lynch and Clark, 2003; Narang et al, 2008), and the mixed quality of the study design in general. In particular, Roberts et al (2006) used sub-therapeutic doses of morphine, which may have limited that study’s ability to test the effects of co-administration. Portenoy et al (2012) noted that the use of fixed dose ranges of cannabinoids may have limited that study’s ability to test the efficacy of cannabinoids for pain, as some patients may have dropped out due to tolerability. Moreover, by discouraging patients from reducing their opioid dose during the study, no opioid-sparing effect could be observed (Portenoy et al, 2012).

This review highlights some important considerations for future studies of cannabinoids. A dose-ranging study with patients with advanced cancer found that only lower doses of cannabinoids demonstrated analgesic effects (Portenoy et al, 2012). In the same study, one in four participants in the high-dose group discontinued treatment. Side effects such as nausea, drowsiness, and dizziness are more frequent with higher doses of cannabinoids (Narang et al, 2008; Portenoy et al, 2012). This suggests that dose range and tolerability are important outcomes to examine and that careful dose titration is essential. Future studies should carefully document adverse effects from concurrent opioid and cannabinoid administration to provide a better understanding of potential harms. One hypothesis to explain why patients reduce their opioid dose with cannabinoid administration is that they experience undesirable psychoactive effects from concurrent use of opioids and cannabinoids. This could be explored in future studies.

Recent observational studies provided further data on a possible opioid-sparing effect. Two studies found 44–64% reductions in self-reported opioid consumption in cohorts of patients with chronic pain who were using cannabis (Boehnke et al, 2016; Haroutounian et al, 2016). These observational studies provide further low-quality evidence supporting an opioid-sparing effect. A further observational study found that in patients with chronic pain who were prescribed opioids, greater pain relief was reported from cannabis than from their other medications (Degenhardt et al, 2015). A single case study also reported reduced requirements for breakthrough pain with oral delta-9-THC administration (Holdcroft et al, 1997). Taken together, these reports support the need for high-quality studies to directly assess the opioid-sparing effect of cannabinoids under controlled conditions.

This review identified some limitations in the literature. The pre-clinical studies examined used a range of animal populations, antinociceptive assays, opioids, and cannabinoids, and often had small numbers of animals per group. This resulted in statistical heterogeneity. Despite this, a large and significant effect was observed in the meta-analysis. No studies examined the opioid-sparing effect of cannabidiol alone, in combination with delta-9-THC outside of a 1 : 1 ratio, or with other cannabinoids. Further, the lack of high-quality studies in humans investigating the opioid-sparing effect means that the evidence for this is largely limited to pre-clinical studies. A funnel plot was produced and did not provide evidence of publication or small study bias; however, due to the small number of studies in the meta-analysis (<10) the interpretation of the funnel plot is limited.

The potential for cannabinoids to reduce opioid dose requirements and extend the duration of effective analgesia should not be understated. The rapid increase in opioid use and opioid-associated mortality is largely attributed use of opioids in chronic pain treatment (Chou et al, 2015; Zedler et al, 2014). Use of lower opioid doses has been recommended (Dowell et al, 2016); however, clinical processes to achieve this reduction are not well defined. Opioid-sparing medications may have enormous clinical relevance by enabling effective pain treatment with lower opioid doses and a potential reduction in opioid-related mortality.

In conclusion, pre-clinical studies support the opioid-sparing effect of delta-9-THC. However, the findings from clinical trials are inconsistent, with some studies found to have important limitations such as a lack of placebo control. An opioid-sparing effect of cannabinoids in chronic pain patients was observed in only one very-low-quality clinical study. These findings provide an early signal that warrants exploration. It remains to be seen if these promising pre-clinical and observational findings can be replicated in large, well-designed clinical studies.

Funding and disclosure

SN is supported by a NHMRC Research Fellowship (#1013803). The National Drug and Alcohol Research Centre at the University of New South Wales is supported by funding from the Australian Government under the Substance Misuse Prevention and Service Improvements Grant Fund. The contents of the published material are solely the responsibility of the authors and do not reflect the funding bodies. MAW has received a grant to his institution from CanniMed. BLF has received speaker fees or consulting fees from Allergan, Mettrum, CCIC, Mylan Pharmaceutical, Pfizer, Ethypharm, Richter Pharmaceuticals, and Lundbeck. He also received salary/grant support from Pfizer and Bioprojet, and in kind support from GW Pharma, Mylan Pharmaceuticals, and Brainsway. SN and NL have been investigators on untied educational grants from Reckitt-Benckiser. SN and MF have been investigators on an untied education grant from Indivior. KEK has previously received a speaker’s honorarium from Pfizer and Mundipharma, in addition to fees from an advisory board and an educational grant from Seqirus. As Director of NDARC, MF notes that the National Drug and Alcohol Research Centre has received untied educational grants from Mundipharma and Indivior. MF took part in a single research advisory board with Indivior in 2014. The remaining authors declare no conflict of interest.