THC Modulates Pain Sensitivity Among Persons Receiving Opioid Agonist Therapy for Opioid Use Disorder: A Within-Subject, Randomized, Placebo-Controlled Laboratory Study

Abstract

The opioid and cannabinoid receptor systems are inextricably linked — overlapping at the anatomical, functional, and behavioral levels. Preclinical studies have reported that cannabinoid and opioid agonists produce synergistic anti-nociceptive effects. Still, there are no experimental data on the effects of cannabinoid agonists among humans who receive opioid agonist therapies for opioid use disorder (OUD). We conducted an experimental study to investigate the acute effects of the delta-9-tetrahydrocannabinol (THC) among persons receiving methadone therapy for OUD. Using a within-subject, crossover, human laboratory design, 25 persons on methadone therapy for OUD (24% women) were randomly assigned to receive single doses of oral doses of THC (10 mg or 20 mg, administered as dronabinol) or placebo, during three separate five-hour test sessions. Measures of experimental and self-reported pain sensitivity, abuse potential, cognitive performance, and physiological effects were collected. Mixed-effects models examined the main effects of THC dose, and interactions between THC (10 mg, 20 mg) and methadone doses (low-dose methadone defined as < 90 mg/day; high-dose defined as > 90 mg/day). Results demonstrated that, for self-reported rather than experimental pain sensitivity measures, 10 mg THC provided greater relief than 20 mg THC, with no substantial evidence of abuse potential, and inconsistent dose-dependent cognitive adverse effects. There was no indication of any interaction between THC and methadone doses. Collectively, these results provide valuable insights for future studies aiming to evaluate the risk-benefit profile of cannabinoids to relieve pain among individuals receiving opioid agonist therapy for OUD, a timely endeavor amidst the opioid crisis.

1. Introduction

The United States remains in the midst of an opioid epidemic, with the toll from opioid overdoses reaching over 68,000 deaths annually ¹. In the United States alone, the annual economic burden of this crisis, in terms of medical, social welfare, and correctional services, is approaching a cumulative cost of $1 trillion, not to mention the incalculable suffering and grief that individuals and families endure ².

In addition to carrying a risk of death by overdose, opioids can induce several neuroadaptations that often make treating pain and opioid use disorder (OUD) challenging ³. First, when used for long periods and at higher doses, opioids induce tolerance to their analgesic effects ⁴, which is sometimes accompanied by a higher sensitivity to pain — a phenomenon known as opioid-induced hyperalgesia ⁵. Notably, among persons with OUD, unrelieved pain is associated with negative clinical outcomes, including impaired social functioning, poor sleep, and OUD treatment dropout ^6–8. Second, opioids cause dose-dependent adverse gastrointestinal, immune, and endocrine effects, which collectively may lead to a lower quality of life and a higher utilization of healthcare services ^9,10. Together, these challenges underscore the need for novel strategies to enhance pain and OUD therapeutics, an initiative recognized by the U.S. Food and Drug Administration (FDA) Congressional Research Service ¹¹. One promising approach is to use adjuvant medications that reduce the opioid dose required to produce a therapeutic response (e.g., analgesia), while mitigating adverse effects, functional impairment, and the risk of respiratory depression ¹². While some opioid analgesic adjuvants exist (e.g., non-steroidal anti-inflammatory drugs), more effective compounds that are suitable for acute co-administration with opioids are urgently needed.

Accumulating preclinical evidence indicates that cannabinoid agonists have opioid-sparing effects (i.e., reducing the analgesic dose of opioids) ^13,14. These effects can be attributed to the considerable neurobiological overlap between the opioid and cannabinoid receptor systems, which share signaling pathways mediating pain and reward ^15,16. Meta-analyses of preclinical studies have shown that when opioids are co-administered with delta-9-tetrahydrocannabinol (THC) — a cannabinoid receptor 1 (CB1-R) partial agonist — the doses of opioids required to produce the same analgesic effect are up to 9.5 times lower ^13,14. Furthermore, preclinical studies have shown that other CB1-R agonists may reduce the development of opioid tolerance ¹⁷, alleviate opioid withdrawal ^18,19, and attenuate opioid self-administration ²⁰. As opioid-induced hypoactivity and respiratory depression are not potentiated by co-administration with cannabinoid agonists, some mechanistic data indicates that their synergistic effects may be selective for anti-nociception ²¹. Thus, if analogous preclinical findings of cannabinoid and opioid agonist synergy were translated to humans, cannabinoid-opioid combinations could hold promise to curb opioid-related harm.

Though the appreciation of opioid-related risks has grown, the medicinal use of cannabinoids has, instead, garnered widespread public acceptance. At the time of writing, 37 states and Washington D.C. have authorized the medicinal use of cannabis and its constituent cannabinoids ²². Pain is the most common qualifying condition, and at least six states have added “opioid use disorder”; “alternative to opioid treatment”; and “all conditions for which opioids could be prescribed to treat” to the list ²². This shift in public policy is dovetailed by early observational data showing that medicinal cannabis laws were associated with fewer opioid overdose deaths at the state level; however, these data cannot speak to causality at the individual level, and there is evidence that this association may have reversed over time ^23,24. Hence, high-quality experimental data to guide policy and clinical decision-making is still lacking.

Thus far, no experimental studies have examined the effects of cannabinoid agonists among persons receiving opioid agonist therapy for OUD. Among healthy persons, randomized, placebo-controlled studies have found that cannabis and THC may increase experimental pain threshold and tolerability ²⁵; and one study reported that cannabis administration reduced the acute analgesic dose of oxycodone ²⁶. Similarly, clinical trials conducted among persons who receive long-term opioid therapy for chronic pain have found evidence of THC-induced analgesia ^27,28. Still, these findings cannot be generalized to persons with OUD who: 1) have a history of using opioids non-medically, and 2) are typically maintained on opioid doses over ten times higher than those of persons receiving long-term opioid therapy doses for pain, and thus have developed more profound opioid neuroadaptations (e.g., analgesic tolerance) ²⁹. Further, some preclinical and human data show that the co-administration of cannabinoid and opioid agonists may increase their reinforcing effects and abuse potential ^26,30–32, a critical concern among persons with OUD, who identify abstinence from non-medical substance use as a critical step towards recovery ³³. Therefore, a rigorous examination of the risks and benefits of administering cannabinoid agonists among persons with OUD receiving opioid agonist therapy is needed.

To address this critical gap, here we report findings of the first experimental human study designed to investigate the acute effects of THC — the main psychoactive component of cannabis — among persons with OUD who are receiving methadone treatment, a first-line opioid agonist therapy for OUD ^34,35. In this within-subject, randomized, placebo-controlled, crossover study, methadone-maintained participants were assigned to receive a single oral dose of THC (10 mg, 20 mg), administered as dronabinol; or placebo, across three randomly ordered test sessions. Pain sensitivity in response to THC administration was measured using experimental assays that have predictive validity for pain therapeutics and self-reported pain responses ^36,37. The subjective drug effects of THC — including measures of abuse potential and opioid withdrawal suppression —, as well as its cognitive and physiological effects, were also assessed using well-validated procedures.

2. Methods

2.1. Participants

Participants were men and women aged between 18 to 70 years old, recruited from opioid treatment programs in the greater New Haven, CT area, using public transportation advertisements, as well as referrals from clinicians in the community. Volunteers who met basic eligibility criteria came to the laboratory for an in-person screening, which included a full psychiatric and medical evaluation. Eligible participants were adherent to a stable dose of methadone for OUD for ≥ 3 weeks; had been previously exposed to cannabis and/or its constituent cannabinoids, by self-report; were physically healthy, as determined by a physical examination, electrocardiogram, and blood chemistry; and capable of providing informed consent in English. Volunteers were excluded if they met DSM-5 criteria for cannabis use disorder and/or substance use disorders (SUD) other than OUD or tobacco use disorder within the last 12 months; had a history of primary psychotic disorders or mood disorders with psychotic features; had any clinically significant abnormalities during the physical examination that might affect safety, study participation, or interpretation of the study results; had any neurological or psychiatric disorder that might affect cognition or the ability to meaningfully participate in the study; were regularly using over-the-counter or prescription medications known to affect pain threshold or pain tolerance, other than methadone; had contraindications for exposure to cold temperatures, such as Raynaud’s phenomenon or hypertension; had a positive pregnancy test; or had experienced serious adverse reactions to cannabis or its constituent cannabinoids. Participants were informed of the study aim to investigate the effects of oral THC on experimental and self-reported pain responses, and that during each test session they would receive one of two doses of THC (10 mg, 20 mg), or placebo. Volunteers were enrolled in the study once they provided informed consent and were compensated for participation. All the study procedures were approved by the Institutional Review Boards (IRB) of the VA Connecticut Healthcare System and Yale University, and were in accordance with the Declaration of Helsinki ³⁸.

2.2. Study Design and Procedures

The study consisted of three five-hour test sessions, all conducted at a dedicated laboratory located at the VA Connecticut Healthcare System Campus in West Haven, CT. Each session began at 8 AM and each were separated by 72 hours to limit carryover effects. During each one of the three test sessions, participants received either a single capsule containing dronabinol (10 mg, 20 mg) or matched placebo. A within-subject, crossover design was used, in which all participants served as their own control and received all three dose conditions in a randomized, counter-balanced order.

Upon arrival at the laboratory, abstinence from alcohol and non-medical substances (including any cannabinoids not administered in the study) was confirmed using the Timeline Followback (TLFB) Method ³⁹, in addition to a breathalyzer and a urine drug screening test. A study physician confirmed that the participants had not received their daily morning dose of methadone in the morning and were, therefore, expected to be at their trough opioid level. Emerging evidence indicates that during this interval between methadone doses, persons with OUD may experience early signs of opioid withdrawal and higher sensitivity to pain ⁴⁰, and thus might be more responsive to THC-induced analgesia.

Participants were instructed not to eat breakfast before sessions, to control for possible effects of food on THC absorption. Participants were also instructed to have their usual amount of nicotine and/or caffeine to avoid confounding effects of withdrawal; they were not, however, allowed to consume nicotine or caffeine once the session started.

Before receiving the study medication, participants completed baseline assessments for heart rate and blood pressure. Following the administration of the study medication assigned for that day, participants underwent serial assessments of pain, subjective drug effects, cognitive performance, and physiological effects, conducted at set time points during the test session (Supplement S1). Participants were instructed to withhold their daily methadone dose until the end of the session. Compliance was confirmed by a study physician who verified the daily methadone dose was brought to the laboratory. Following the 180-minute session timepoint (3.5 hours after arrival in the laboratory), participants received their daily methadone dose and a standardized, low-fat meal. Following all test procedures, approximately five hours after arrival in the laboratory, a study physician conducted a mental status exam and sobriety tests to ensure that participants were safe for discharge. All participants agreed, via written consent, not to drive for at least 24 hours following each session and received compensation to assist with transportation costs to leave the laboratory safely.

2.3. Study Medication

THC was administered as dronabinol, an orally active, Schedule III, synthetic cannabinoid agonist that is FDA-approved to treat HIV-associated anorexia and chemotherapy-induced nausea and emesis. While the analgesic and psychoactive effects of smoked cannabis have their onset within a few minutes, dronabinol produces a gradual increase in plasma levels, providing prolonged analgesia with a lower abuse potential ⁴¹. Specifically, the bioavailability of dronabinol is 20-25%, with peak effects noted between 30 to 180 minutes ⁴². The doses of dronabinol (10 mg and 20 mg) administered in this study were chosen based on prior evidence of dronabinol-induced analgesia among persons receiving long-term opioid therapy within the context of a randomized, placebo-controlled clinical trial ²⁷. Medication capsules (size 00 opaque capsules with lactose filter) containing dronabinol (10 mg, 20 mg) or placebo were prepared by the VA Connecticut Healthcare System Research Pharmacy, on the day of each test session.

3. Measures

3.1. Pain Sensitivity

The Cold Pressor Test (CPT) and the McGill Pain Questionnaire (MPQ) were selected as experimental and self-reported measures of pain sensitivity, respectively, given their known responsiveness to THC-induced analgesia ^43,44. To capture the peak analgesic effects of THC, pain responses were measured at 90 minutes and 180 minutes following the administration of the study medication.

The CPT is a measure of evoked pain that has shown predictive validity for detecting the analgesic effects of several medications used to treat chronic pain ^36,37. In this test, participants immersed their forearm into a cold-water bath (4°C / 39.2°F) and were instructed to 1) report the time the cold sensation become a painful sensation (pain threshold) and, 2) withdraw their arm from the water when the pain becomes intolerable (pain tolerance), both measured in seconds. Staff administering the CPT was the same sex as the study participant.

Immediately following the CPT, a 15-item version of the MPQ was administered, to assess the sensory and affective dimensions of the pain experience. Prompted by sensory (e.g., “throbbing”, “shooting”) or affective (e.g., “punishing”, or “fearful”) descriptors, participants were instructed to describe their pain experience, by selecting among a Likert-scale ranging from “not at all” (score = 0) to “severe” (score = 3). Scores were summed for sensory items (ranging from 0 to 33); affective items (ranging from 0 to 12); and all items, yielding a total pain score (ranging from 0 to 45).

3.2. Subjective Drug Effects

The subjective effects of THC were assessed using the Drug Effects Questionnaire (DEQ) at baseline and every 30 minutes following the administration of study medication, to measure abuse potential ⁴⁵. The DEQ has been show to sensitive to the effects of THC and is composed of 10 items measured on a 0 to 100 mm visual analogue scale, anchored in the statements “not at all” and “extremely”, which are subsequently converted to a numeric score ⁴⁶. DEQ ratings were grouped into three categories: (1) Stimulatory effects: the average of “feel stimulated”, “feel the drug strength”, and “feel high”; (2) Pleasurable effects: the average of “like the drug effect”, “feel good drug effects”, and “want more of the drug received”; and (3) Aversive effects: the average of “feel anxious”, “feel down”, and “feel bad drug effects”. Each DEQ subscale ranges from 0 to 100 mm.

Subjective symptoms of opioid withdrawal were also assessed at baseline and every 30 minutes following administration of the study medication using the Subjective Opioid Withdrawal Scale (SOWS) ⁴⁷. The SOWS has 16 items, each representing a specific opioid withdrawal symptom, such as “anxiety”, “irritability”, “restlessness”, and “muscle aches” ⁴⁷. Participants rate the intensity of each symptom on a scale from 0, “not at all”, to 4, “extremely”, with higher scores indicating more severe withdrawal symptoms, yielding a score from 0 to 64 ⁴⁷.

3.3. Cognitive Performance

Participants’ cognitive performance was assessed using the Continuous Performance Test and the Hopkins Verbal Learning Test (HVLT) ^48,49. Both the continuous performance test and the HVLT were administered once per test session, at timepoint 150 minutes after the study medication administration.

The continuous performance test is used to assess sustained attention and working memory and is sensitive to the effects of THC ⁵⁰. The continuous performance test requires participants to press the left button of an external mouse when the displayed number (i.e., 0–9) matches the number that immediately preceded it. If the displayed number is not the same as the preceding number, the participant is asked to press the right button of the external mouse. Of the 80 trials, 50% display a number that matches the preceding number, and 50% display a number that does not match the preceding number. Each number is displayed for 2500 ms, and the participant has 3000 ms to respond before the next number is presented. Consistent with best practices ⁴⁸, continuous performance test administrations that had low accuracy (<56%) were excluded from the final analysis. The main outcome measures for the continuous performance test were the throughput score and percent correct responses. The throughput score accounts for task performance speed and accuracy ⁴⁸.

The HVLT assesses verbal learning. In this test, participants are read a list of 12 words aloud and then immediately asked to repeat as many words as they can remember in any order, over a series of three trials. Twenty minutes later, participants are once again asked to repeat as many of the words listed as they can remember. The main outcome measures for the HVLT were total immediate recall score, represented by the sum of the first three trials, thus ranging from 0 to 36; and the delayed recall total score, represented by the words recalled after a 20-minute delay, thus ranging from 0 to 12. A wealth of evidence shows that these verbal learning measures are sensitive to the acute effects of THC ⁵¹.

3.4. Physiological Effects

Blood pressure and heart rate were operationalized as indexes of the physiological effects of THC. Systolic and diastolic blood pressure and heart rate were measured at baseline and every 30 minutes after the administration of study medication (Supplement S1).

4. Statistical Analysis

Each dependent variable was analyzed using a mixed-effects model with the within-subject factors of THC dose (10 mg, 20 mg) and the interaction between THC dose and time. Random effects for participant and structured variance–covariance matrices for the repeated measures on each participant within-session were used to account for the correlation structure of the data. The best-fitting structure for each outcome was selected based on the Schwartz–Bayesian Information Criterion. Examination of model residuals confirmed model fit. Least square means and standard errors were calculated to describe the patterns of means for each outcome. Tests of effect sizes were used to describe significant interactions. Least square means and standard errors were estimated and compared post-hoc to explain significant main and interactive effects. All analyses were conducted at the two-sided alpha=.05 threshold and conducted using SAS, version 9.4 (Cary, NC).

Pairwise comparisons of least square means were used to describe the significant main and interactive effects, focusing on the main effect of THC dose and interactions between THC dose and time. Additional analyses were conducted at peak timepoint (180 minutes after THC), including the daily methadone dose (low vs. high) as between-subject factor. Specifically, we explored the interaction between the within-subject factor THC dose (10 mg, 20 mg) and the between-subject factor methadone dose, with low-dose methadone defined as lower than 90 mg/day and high-dose methadone defined as higher than 90 mg/day ^52,53.

5. Results

5.1. Demographic Characteristics

A total of 27 participants (seven women) were enrolled in the study and randomized to receive the study medication. A total of 20 participants completed all study procedures. Five participants (all men) completed two out of the three test sessions, and thus provided partial data. Among these five participants, three experienced an adverse effect possibly related to the study medication (two developed anxiety-like symptoms and one developed acute hypertension); one participant was discharged due to non-medical substance use during the study, evidenced by urine drug screens; and one participant voluntarily withdrew participation due to personal reasons unrelated to the study medication. Two participants (one male and one female) only completed one of the three test sessions and were thus excluded from the analysis. One experienced an adverse effect possibly related to the study medication and one was discharged due to personal reasons unrelated to the study medication.

A total of 25 participants (six women) were included in the analyses. Participants were aged 47.4 ± 12.3 years old. The average methadone dose was 94.6 ± 35.1 mg/day (1159.2 ± 280.8 morphine milligram equivalents per day). Approximately 44% of participants were receiving low-dose methadone (63.3 ± 22.7 mg/day), and 56% were receiving high-dose methadone (119.3 ± 19.8 mg/day). As the methadone dose did not impact the study outcomes, the low-dose and high-dose methadone groups were merged, and the data are presented as a function of THC dose. Table 1 shows the demographic characteristics of the study participants and Table 2 shows the main results. The mean number of days between test sessions was 7.34 (SD = 6.98). Supplement S2 shows the Consolidated Standards of Reporting Trials (CONSORT) Diagram.

Table 1.

Demographic Characteristics

Demographics (n= 25)	N (%)	Mean (SD)
Sex
Male	19 (76%)
Female	6 (24%)
Race/Ethnicity
White	17 (68%)
African-American or Black	6 (24%)
Hispanic or Latino	2 (8%)
Age		47.36 (12.26)
Methadone Dose (mg)
Daily Dose		94.64 (35.11)
High-dose methadone (>90mg)	14 (56%)	119.29 (19.79)
Low-dose methadone (<90mg)	11 (44%)	63.27 (22.72)
Cannabis Use
Age of first use		17.31 (1.96)
Years of use		29 (12.15)
Average joints/week		0.58 (0.89)
Drug Use in Past 28 Days (days)
Cannabis	6 (24%)	8.50 (5.09)
Tobacco	12 (48%)	26.83 (4.04)
Alcohol	3 (12%)	1.33 (0.58)
Cocaine	4 (16%)	1.25 (0.50)
Opioids	2 (8%)	1.00 (0.00)
Sedatives	4 (16%)	3.75 (5.50)
Stimulants	0 (0%)	0.00 (0.00)

Drug use in the past 28 days was measured using the Timeline Followback (TLFB) method.

Table 2.

Summary of Outcomes

	10mg THC	20mg THC	Placebo
	Mean (SE)	Mean (SE)	Mean (SE)
Pain Sensitivity
CPT
Pain Threshold (sec)	20.23 (3.36)	18.38 (3.30)	15.07 (3.32)
Pain Tolerance (sec)	66.69 (14.65)	60.15 (14.22)	63.73 (13.49)
MPQ
Sensory Pain	9.97 (1.47)	10.60 (1.43)	12.99 (1.44)
Affective Pain	2. 34 (0.56)	2.20 (0.55)	2.56 (0.55)
Total Pain	12.39 (1.91)	12.81 (1.88)	15.54 (1.88)
Subjective Drug Effects
DEQ
Stimulatory Effects	10.75 (2.25)	14.07 (2.19)	7.63 (2.19)
Pleasurable Effects	9.23 (2.39)	10.04 (2.28)	6.88 (2.28)
Aversive Effects	6.30 (2.02)	6.41 (1.99)	5.67 (1.99)
Cognitive Performance
HVLT
Immediate Recall	18.86(1.29)	17.05 (1.270)	20.00 (1.270)
Delayed Recall	5.47(0.63)	4.52 (0.63)	6.03 (0.63)
Continuous Performance Test
Throughput Score	70.64 (4.12)	69.90 (4.04)	72.11 (4.06)
Percent Correct Responses	90.77 (2.33)	90.53 (2.37)	95.41 (1.144)
Physiological Effects
Systolic BP	123.62 (2.35)	122.53 (2.27)	126.02 (2.29)
Diastolic BP	74.59 (2.01)	73.77 (1.95)	74.36 (1.97)
Heart Rate (BPM)	72.67 (2.17)	74.61 (2.10)	70.25 (2.12)

Abbreviations: SE: Standard Error; CPT: Cold Pressor Test; MPQ: McGill Pain Questionnaire; DEQ: Drug Effects Questionnaire; HVLT: Hopkins Verbal Learning Test; BP: Blood Pressure, BPM: Beats Per Minute; THC: Delta-9-tetrahydrocannabinol, administered as dronabinol. Means are least square means derived from the linear mixed models.

5.2. Pain Sensitivity

For the CPT pain threshold (Figure 1A), there was no significant main effect of THC dose nor THC dose-by-time interaction. Similarly, for CPT pain tolerance (Figure 1B), significant main effects of THC dose nor THC dose-by-time interaction were not observed. Notably, CPT pain tolerance was highest at peak effects timepoint (180 minutes) with 10 mg THC (d’ = 0.3), but this finding was not significantly different from placebo or from 20 mg THC.

Figure 1. Effects of Delta-9-tetrahydocannabinol (THC) on Pain Sensitivity Indexed by The Cold Pressor Test (CPT).

Mean scores for (A) pain threshold and (B) pain tolerance of the CPT, at 90 minutes and 180 minutes, in response to oral THC (10 mg, 20 mg) and placebo. The maximum duration of the CPT is of 180 seconds. For details, please see section 5.2 of the manuscript.

For the MPQ total pain score, which assesses the overall severity of the pain experience, a significant main effect of THC dose (F(2,104) = 5.33, p = .006) was observed (Figure 2A), owing to lower total pain reported with 10 mg THC (t(2,104) = −2.94, p = .004, d’ = 0.65) and 20 mg THC (t(2,104) = −2.69, p = .03, d’ = 0.08) than with placebo. Post hoc analyses revealed that lower MPQ total pain scores observed with THC administration were primarily explained by a reduction of sensory pain (F(2,104) = 5.92, p =.004) (Figure 2B), rather than due to a reduction of affective pain (F(2,104) = .52, p =.59) (Figure 2C).

Figure 2. Effects of Delta-9-tetrahydocannabinol (THC) on Pain Sensitivity Indexed by the McGill Pain Questionnaire (MPQ).

Mean scores for (A) total pain (B) sensory pain, and (C) affective pain, at 90 minutes and 180 minutes, indexed by the MPQ, in response to oral THC (10 mg, 20 mg) and placebo. The MPQ total pain scores range from 0 to 45. The MPQ sensory items range from 0 to 33 and the affective items range from 0 to 12. For details, please see section 5.2 of the manuscript. *p< .05 for pairwise comparisons of the THC conditions with the placebo condition.

5.3. Subjective Drug Effects

With respect to the abuse potential of THC, for the Stimulatory Effects of the DEQ (Figure 3A) there was a main effect of THC dose (F(2,463) = 4.05, p = .01) and a THC dose-by-time interaction (F(14,463) = 2.33, p = .004). Post hoc analyses showed that both 10 mg THC (F(7,463) = 3.36, p = .001, d’ = 0.4) and 20 mg THC (F(7,463) = 4.38, p = .0001, d’ = 0.5) produced greater Stimulatory Effects ratings than placebo. Notably, the greatest differences between THC and placebo were observed between the timepoints 150 min and 210 minutes post-medication. For the DEQ Pleasurable Effects (Figure 3B), neither the main effect of THC dose nor the THC dose-by-time interaction were significant. Similarly, for the DEQ Aversive Effects (Figure 3C), no significant main effects of THC dose or THC dose by time interaction were observed. It is worth noting that the increases in DEQ produced by active doses of THC measured approximately 18 mm in a 100 mm scale, thus suggesting limited abuse potential ⁴⁶.

Figure 3. Subjective Effects of Delta-9-tetrahydocannabinol (THC).

Mean scores for (A) stimulatory effects (B) pleasurable effects, and (C) aversive effects, indexed by the Drug Effects Questionnaire (DEQ); and (D) opioid withdrawal severity, indexed by the Subjective Opioid Withdrawal Scale (SOWS), in response to oral THC (10 mg, 20 mg) and placebo. The DEQ measures range from 0 to 100 mm, and the SOWS scores range from 0 to 64. Error bars reflect standard error of the mean. For details, please see section 5.3 of the manuscript. *p< 0.05 for pairwise comparisons of the THC conditions with the placebo condition.

For subjective symptoms of opioid withdrawal, indexed by the SOWS, neither the main effect of THC dose nor the interaction between THC dose-and-time were significant (Figure 3D).

5.4. Cognitive Performance

For sustained attention and working memory performance, indexed by the continuous performance test throughput score (Figure 4A), we did not observe a main effect of THC dose; however, for continuous performance test percent correct responses, there was a main effect of THC dose (F(2,36) = 6.68, p = .003) (Figure 4B), owing to significantly lower percent correct responses with both 10 mg THC (t(2,36) = −3.36, p = .0019) and 20 mg THC (t(2,36) = −3.44, p = .0015) than with placebo.

Figure 4. Effects of Delta-9-tetrahydocannabinol (THC) on Cognitive Performance.

Mean scores for (A and B) sustained attention, indexed by the Continuous Performance Test (ContPT) throughput score and percent correct responses, respectively; and (C and D) verbal learning, indexed by the Hopkins Verbal Learning Test (HVLT) total/immediate and delayed recall, respectively, in response to oral THC (10 mg, 20 mg) and placebo. The continuous performance test throughput scores range from 0 to 100. The HVLT immediate recall scores range from 0 to 46 and the HVLT delayed recalls scores range from 0 to 12. For details, please see section 5.4 of the manuscript. *p< 0.05 for pairwise comparisons of the THC conditions with the placebo condition.

For verbal learning, a significant THC dose effect was observed for delayed recall (F(2,39) = 3.72, p = .03, d’ = 0.4) (Figure 4D), with 20 mg producing verbal learning deficits relative to placebo. Similar THC effects on total immediate recall were observed (d’ =0.4), although the dose effect did not reach significance (Figure 4C).

5.5. Physiological Effects

For systolic blood pressure, neither the main effect of THC dose nor THC dose-by-time interaction were significant. Similarly, for diastolic blood pressure, neither the main effect of THC dose nor THC dose-by time-interaction were significant. Finally, for heart rate, we observed increases under THC, but the dose effect did not reach statistical significance. There were no significant THC dose-by-time interactions.

6. Discussion

We employed a carefully controlled experimental design to measure the potential therapeutic and adverse effects of THC among persons receiving opioid agonist therapy for OUD — a logical next step towards estimating the risk/benefit ratio of therapeutic cannabinoid use in this population. This study had several notable findings. First, both 10 mg and 20 mg THC alleviated pain on self-reported measures, but not on the experimental pain assay; notably, the effects of 10 mg THC on self-reported measures of pain were superior to those of 20 mg THC. Second, our results indicated that, among persons receiving methadone therapy for OUD, THC may preferentially modulate sensory, as opposed to affective pain processes. Third, we found limited evidence of dose-dependent abuse potential, and inconsistent cognitive adverse effects of THC. Fourth, we found no evidence that the daily methadone dose influences the acute effects of THC. Collectively, these findings provide key insights into the potential therapeutic and adverse effects of cannabinoids among persons receiving opioid agonist therapy, and offer methodological and mechanistic implications for future research.

6.3. Implications for Understanding the Analgesic Effects of THC

Thus far, human laboratory studies investigating the effects of cannabis and cannabinoid agonists administered alone (e.g., THC, dronabinol, or nabilone) to healthy persons have found mixed experimental pain results ²⁵. Similarly, some but not all prior human laboratory studies and clinical trials have shown that the addition of THC-based substances — dronabinol or cannabis — may enhance the analgesic effects of either acute or chronically administered opioids ^26,31,54. Among the five human laboratory studies co-administering cannabis/THC and opioids to healthy (i.e., non-opioid dependent) persons, three found evidence of analgesic effects of cannabinoids on experimental pain outcomes ^26,31,54 and two did not ^55,56. Only two randomized, placebo-controlled human laboratory studies administered cannabis or THC to persons who were already opioid-dependent (mean daily long-term opioid therapy dose of 68 to 154 morphine milligram equivalents per day) ^27,28, and both studies reported modest THC-induced analgesia on clinical, rather than experimental pain outcomes ^27,28. Taken together, the preponderance of the evidence from human experimental and clinical pain studies indicates that the analgesic efficacy of cannabinoid agonists may vary as a function of acute versus chronic opioid exposure at baseline. Our study extends those findings, by 1) assessing the effects of THC on sensory and affective aspects of the pain experience, among persons receiving opioid agonist therapy for OUD, a unique clinical population due to pre-existing opioid addiction and high levels of opioid tolerance; and by 2) examining dose-dependent analgesic effects of THC, as well as potential dose interactions between THC and methadone, a full mu-opioid receptor (MOR) agonist and first-line pharmacotherapy for OUD. Notably, the average methadone dose in our study corresponded to 1159.2 ± 280.8 morphine milligram equivalents per day. This is over 13 times the level defined by the U.S. Center for Disease Control and Prevention as high-dose opioid therapy ⁵⁷. Mechanistic evidence suggests that chronic and high doses of MOR agonists, such as the therapeutic doses of methadone in our sample, may result in desensitization and downregulation of CB1-R ⁵⁸. This has potential implications for the regulation of nociception, anxiety, and negative affective states ⁵⁹. Considering the varying different levels of sensitivity and availability of CB1-R receptors across healthy persons and distinct clinical populations, THC may primarily impact nociceptive sensory or affect-related pain processes. Interestingly, we observed that 10 mg THC was associated with lower self-reported measures of pain than 20 mg THC. This observation aligns with convergent data showing biphasic effects of THC on affective states, with lower doses of THC typically having an anxiolytic effect, while higher doses tend to be anxiogenic ^60,61. Extensive research has demonstrated a strong link between anxiety, negative affective states, and the experience of pain ^62,63. Moreover, cannabis and THC have been shown to modulate these states ²⁵. However, there is also evidence that cannabis and THC may have a limited window of analgesic efficacy. High-THC content cannabis and doses of 20 mg of oral THC have been reported to produce hyperalgesia, as opposed to analgesia, in experimental models of pain ^64,65. Therefore, these observations warrant further investigation to elucidate the complex relationship between THC, pain, and affective states among persons with OUD.

6.2. Implications for Understanding the Subjective Drug Effects of THC

In terms of abuse potential, THC elicited mild stimulatory effects that were dose-dependent , with measurements of approximately 18 mm in a 100 mm scale; however, there was no indication of pleasurable effects observed. In contrast, opioids have been shown to produce robust and consistent increases in measures of subjective drug effects that are indicative of abuse potential, with readings exceeding >80 mm in a 100 mm scale, especially among persons who use opioids non-medically ⁶⁶. Thus, these findings lend support to the notion that THC, when administered orally, has a low potential for abuse in this population.

These findings are consistent with those from prior human laboratory studies examining the abuse potential of opioid-cannabinoid combinations ^26,28,31,56. Notably, most prior human laboratory studies that assessed the subjective effects of co-administered cannabis/THC and opioids were conducted among healthy (i.e., non-opioid-dependent) persons ^26,31,56. Findings from these studies indicate that THC may produce mild increases in measures of subjective drug effects that are indicative of abuse potential.

Only one prior study measured the abuse potential of cannabis among opioid-dependent persons — rather than healthy participants — and it found an increase in subjective ratings of “high”, although without placebo control ²⁸. These results are also in contrast with conflicting preclinical data showing that CB1-R agonists increase ³⁰; do not change ^67,68; or reduce the reinforcement induced by MOR agonists ²⁰.

It is worth noting that participants in our study did not use cannabis regularly (Table 1). Since we found no evidence of THC by methadone interactions, it appears that persons receiving even high doses of methadone remain sensitive to small dose-dependent subjective effects of THC, although the clinical significance of these findings remains uncertain. Finally, although we did not observe a THC-induced alleviation of opioid withdrawal, the methadone dose was only held for 3.5 hours in this outpatient study. While this period was long enough to produce discomfort, longer periods without opioids may be required to test opioid withdrawal-alleviating effects of THC in the inpatient setting, consistent with prior human research ^69–71.

6.3. Implications for Understanding the Cognitive and Physiological Effects of THC

We observed evidence of acute, THC-induced attention and verbal learning deficits. These findings are consistent with a wealth of data underscoring acute cognitive adverse effects induced by cannabis and CB1-R agonists in humans ⁷². Our findings suggest that the acute effects of THC on cognitive performance in this population are dose-dependent, as lower cognitive performance was consistently linked to higher doses of THC, but not to lower doses.

We did not find any effects of THC dose on heart rate and blood pressure. The doses of oral THC administered typically do not evoke a robust response in blood pressure. They may, however, induce tachycardia in healthy participants. Conversely, opioid agonists have the potential to induce bradycardia ⁷³, consequently inhibiting substantial increases in heart rate. Hence, these physiological effects of THC could be attributed to the distinctive characteristics of our sample.

6.4. Methodological and Mechanistic Implications

The methods here described may serve as model to examine the potential therapeutic and adverse effects of cannabinoids — and other non-opioid analgesics — among persons receiving opioid agonist therapy for OUD. Towards that end, we combined multi-dimensional outcome measures — including measures of pain sensitivity, abuse potential, opioid withdrawal, cognitive performance, and physiological outcomes — in accordance with the growing consensus to standardize the study of analgesics that have addictive potential ^74,75. Both cannabinoid and opioid receptors are G-protein coupled receptors with downstream effects of adenylyl cyclase activity, calcium channel activation, and downstream neurotransmitter release ¹⁵. Further, these receptors are densely co-localized in brain regions that mediate pain, reward, and cognition — such as the periaqueductal gray (PAG), locus coeruleus, and higher-level cortices ¹⁵. Future mechanistic studies should examine how opioids with different mechanisms of action (e.g., various levels of selectivity, intrinsic efficacy, affinity, and potency) at MOR modulate the antinociceptive, reinforcing, and cognitive effects of cannabinoids. Given mechanistic data showing that different opioid ligands (e.g., partial vs. full MOR agonists) induce different levels of MOR desensitization and downregulation, the impact of CB1-R agonists against the backdrop of buprenorphine, a partial MOR agonist also used to treat OUD, warrants future research. Likewise, the presence and severity of chronic pain at baseline and the duration of opioid exposure may impact pain sensing and regulation systems, thereby influencing cannabinoid-induced analgesia. Although experimental data from healthy persons suggest that inhaled cannabis and oral THC produce a similar magnitude of analgesia ⁴¹, further studies should examine the risk/benefit ratio of inhaled cannabinoids among persons with OUD, given the prevalence of inhaled consumption. Finally, preliminary data suggests that cannabis produces more robust analgesia among healthy men than healthy women ⁷⁶, requiring adequately powered studies to examine sex-differences in THC-induced analgesia among persons with OUD.

6.5. Limitations

The current findings provide evidence of acute analgesic effects of THC among persons receiving opioid agonist therapy for OUD; however, these results should be interpreted in the context of limitations. The safety of our study was enhanced by selecting participants receiving a stable dose of methadone and without cannabis use disorder or other (non-tobacco) substance use disorders, although a subset of our sample had exposure to cannabis and tobacco within the 28 days prior to the study sessions, which could influence the acute effects of THC. These safe eligibility criteria may limit the generalizability of our findings for persons undergoing changes in methadone dose (e.g., during induction onto or taper), or for persons with OUD who have other substance use disorders. Because the safety and tolerability of cannabinoids depends on past experience and other drug use, understanding the risk/benefit ratio of cannabinoids among different sub-populations is important. Another consideration is that we administered THC orally, which bypasses respiratory risks associated with the smoked or vaporized routes, but at the cost of lower bioavailability. Future research should explore the risk/benefit ratio of other routes of THC administration in this population. An additional limitation of our study was the low number of women participants, which limited our ability to assess sex-dependent effects of THC ⁷⁶. Finally, given the exploratory nature of this study, adjustments for multiple comparisons were not made, and such adjustments could influence the overall significance of the effects.

7. Conclusions

Cannabinoids have been increasingly viewed as a therapeutic option to enhance the therapeutic effects of opioids and to mitigate opioid-related adverse effects. We have experimentally examined the acute effects of THC among receiving persons opioid agonist therapy for OUD. Our results show that THC-induced analgesia may occur, among this extremely opioid-tolerant population, albeit at the expense of limited and dose-dependent abuse potential and cognitive adverse effects. Collectively, these data underscore the need for further double-blind, placebo-controlled studies, to characterize the effects of cannabinoids on multiple aspects of the pain experience, as well as on adverse effects, among this clinical population. The tradeoff between therapeutic benefit and adverse effects may vary not only across various phases of OUD treatment (e.g., induction, maintenance, and taper), but also across different medications for OUD (e.g., methadone and buprenorphine). In summary, further well-controlled research is necessary to ascertain the dose-dependent risk/benefit ratio of cannabinoids among persons receiving various pharmacotherapies for OUD — an urgent public health need, given the rapidly changing attitudes towards cannabinoid use amidst the current opioid crisis.

Data Availability Statement

The data supporting the findings of this study are available upon reasonable request from the corresponding author. These data are not publicly available due to privacy or ethical restrictions.