Effect of Medical Marijuana Card Ownership on Pain, Insomnia, and Affective Disorder Symptoms in Adults

This randomized clinical trial examines the risks and benefits of a medical marijuana card, including the development of cannabis use disorder and improvement of insomnia symptoms.


Introduction
Despite inconclusive evidence of its efficacy 1 and little information on its risk, medical cannabis has surged in popularity.As of December 2021, approximately 12 countries, including Canada, the United Kingdom, and Australia, and 36 US states and the District of Columbia have commercialized cannabis for medical use, making it accessible through the use of a medical marijuana card for myriad health conditions.Given the increasing prevalence of cannabis use for medical concerns, well-designed studies are needed to assess the effect of cannabis product use on target symptoms and associated adverse medical and psychiatric events, particularly the development of cannabis use disorder (CUD).
Cannabis has been reported to improve pain, sleep, and anxiety and depressive symptoms 2 and is commonly sought for these concerns. 3However, according to national data, 3 in 10 US adults who use cannabis develop CUD, with 23% developing severe CUD 4 and often with a tolerance to delta-9tetrahydrocannabinol (THC) and withdrawal symptoms. 5,6Data are lacking on whether the rates of addiction in adults with a medical marijuana card are similar to the rates in those who use cannabis for recreational purposes.][9][10] Thus, it is imperative to better understand both the benefits and potential risks of cannabis use for medical concerns in the current regulatory environment.
We conducted a randomized clinical trial (RCT) to evaluate the effect of obtaining a medical marijuana card on target clinical and CUD symptoms in adults with a chief concern of chronic pain, insomnia, or anxiety or depressive symptoms.Participants were randomized to acquire a medical marijuana card immediately or to be placed on a waiting list to procure a card.We hypothesized modest improvements in pain and insomnia symptoms along with worsened CUD and depressive symptoms over 12 weeks in participants in the immediate card acquisition group.

Methods
This pragmatic, single-site, single-blind RCT was conducted in the Greater Boston area from July 1, 2017, to July 31, 2020, approximately 2 years after medical cannabis dispensaries began operating in Massachusetts.Study procedures (Supplement 1) were approved by the Mass General Brigham Human Research Committee.Written informed consent was obtained from all study participants, and they received financial compensation for participation.We followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline. 11rticipants were recruited from clinical sites (eg, local Massachusetts General Hospital clinics) and from the community.Eligible participants were aged 18 to 65 years who sought medical marijuana to improve pain, insomnia, and anxiety or depressive symptoms.Daily cannabis use, CUD diagnosis at screening or baseline, cancer, psychosis, and current substance use disorders (except for mild or moderate alcohol use disorder and nicotine use disorder) were the criteria for exclusion.
The eMethods in Supplement 2 provides detailed inclusion criteria.

Randomization and Masking
Participants were randomized to either the immediate card acquisition group or to the delayed card acquisition group (Figure 1).In the immediate card acquisition group, participants were allowed to obtain a card immediately.In the delayed card acquisition group, participants were asked to wait 12 weeks to obtain a medical marijuana card.Participants from both groups could choose their cannabis products, dose, and frequency of use, thus allowing for a pragmatic evaluation of the effect of a medical marijuana card within the system in place for physician recommendation and product regulation and distribution.
Randomization was stratified by sex, age (18-25 vs 26-65 years), and primary medical concern (pain, insomnia, or anxiety or depression).Because we expected that financial and logistical constraints might limit the procurement of a medical marijuana card, increasing the dropout rate in the immediate card acquisition group between randomization and the start of study procedures, we randomized participants 2:1 (immediate card acquisition group to delayed card acquisition group) in each stratum to generate groups that were approximately equal in size for analysis.Randomization was accomplished using a computer-generated random number sequence, which was created and executed by an independent statistician.Blinding of participants after group randomization was not possible because of the study design.Rater blinding is discussed in the eMethods in Supplement 2.

Procedures and Measures
Sociodemographic information, medical history, and psychiatric diagnoses were collected at screening, before randomization, using the Mini International Neuropsychiatric Interview 12 (Table 1).
Participants in the immediate card acquisition group then obtained their cards, and all participants returned for in-person visits at baseline and at weeks 2, 4, and 12 after randomization as well as participated in a telephone visit at week 8 for an assessment of adverse events (AEs).Participants in the immediate card acquisition group were responsible for arranging for and paying the costs of obtaining a medical marijuana card and cannabis products.The trial did not provide or pay for the medical marijuana cards or cannabis products used by the participants.
Verification of compliance within the randomization group was self-reported by participants.All participants in the immediate card acquisition group reported obtaining a card before the baseline visit, and all participants in the delayed card acquisition group agreed to wait 12 weeks to procure a card.Quantity and frequency of cannabis use; sleep quality; and depression, anxiety, and pain symptoms were reported and assessed at every visit via interviews and daily via smartphone diaries.
Participants could continue their ongoing medical or psychiatric care during the trial.

Primary, Secondary, and Exploratory Outcomes
Outcomes were assessed at baseline and at weeks 2, 4, and 12 except for the cognitive and the Short-Form Health Survey (SF-12) measures, 13 15 ; (4) pain severity, which was assessed using the severity subscale of the Brief Pain Inventory (score range: 0-10, with 10 being the worst imaginable pain) 16 ; and (5) insomnia symptoms, which were assessed using the Athens Insomnia Scale (score range: 0-24, with higher scores indicating more severe sleep difficulties). 17condary outcomes included physical and mental (assessed using the SF-12 Physical and Mental scales; score range: 0-100; using T-scores, higher scores indicate better physical health and mental health functioning) 13 as well as cognitive (assessed using the Cambridge Neuropsychological Test Automated Battery [CANTAB]) measures. 18 Scale; score range: 0-52, with higher scores indicating greater pain catastrophizing), 21 perceived stress (assessed with the Perceived Stress Scale; score range: 0-40, with higher scores indicating greater perceived stress), 22 suicidal thoughts (assessed with the Concise Health Risk Tracking scale; score range: 12-60, with higher scores indicating more suicidal thoughts), 23 and illness severity and improvement (assessed with the Clinical Global Impression [CGI] 14 Severity subscale [score range: 1-7, with the highest scores indicating greatest severity of illness] and the Improvement subscale [score range: -3 to 3, with higher scores indicating worse-than-baseline condition and negative scores indicating improvement]). 24Cannabis use at each visit was assessed on a 7-point Likert scale (score range: 1-7, with higher scores indicating greater frequency of use).

Urinalysis and Adverse Events
Urine samples were collected from participants at each study visit and then shipped on dry ice to the University of Colorado Department of Anesthesiology, where the samples were analyzed for cannabinoids using high-performance liquid chromatography with tandem mass spectrometry. 25is assay quantified THC, cannabidiol (CBD), primary metabolites, and 15 other cannabinoids. 26 identified AEs via an open-ended question: "Since the last time we saw/spoke to you, have you experienced any medical events such as illness or injury, or worsening symptoms?"Participants who reported any substantial worsening of their psychiatric condition underwent a psychiatric evaluation by a mental health professional.Cannabis use disorder symptoms were a primary

Sample Size and Power
We aimed to recruit 200 participants.To identify a difference in onset of CUD symptoms in the immediate card acquisition group vs the delayed card acquisition group, assuming that the mean number of symptoms was 0.4 (corresponding to 20% of participants developing CUD) in the immediate card acquisition group vs 0.1 in the delayed card acquisition group (corresponding to 5%

JAMA Network Open | Psychiatry
Effect of Medical Marijuana Card Ownership on Pain, Insomnia, and Affective Disorder Symptoms of participants developing CUD) at a 2-sided α = .05significance level, we estimated a power of 85%.
To identify differences in pain, insomnia, and anxiety or depressive symptoms, assuming that a clinically significant effect would be a 30% reduction in the presenting medical symptom in the immediate card acquisition group and a 5% reduction in the delayed card acquisition group at a 2-sided α = .05significance level based on effect sizes in the literature, [27][28][29] we estimated a power of 84% for pain, 90% for insomnia, and 84% for anxiety or depressive symptoms, with sample sizes of approximately 33 in each subgroup.The eMethods in Supplement 2 include the calculation of sample size and power.

Statistical Analysis
All participants who completed a baseline assessment and at least 1 postbaseline visit were included in the analysis.We used an evaluable population approach for the statistical analyses that included all participants in the delayed card acquisition group and all participants in the immediate card educational level, baseline level of cannabis use, primary concern, and symptom severity) were different in the immediate card acquisition group among participants who obtained a card vs those who did not.To test whether the probability of obtaining a card biased participants in both randomization groups, we calculated propensity scores for all participants using the logistic regression model (model run in the immediate card acquisition group and applied to both groups).
We entered these weightings in the analyses of primary outcomes using generalized estimating equations by the inverse of this propensity score.
Second, because all participants could choose a variety of cannabis products with input from licensed medical marijuana dispensaries or elsewhere, we conducted a modified per protocol sensitivity analysis that compared the primary outcomes in participants in the immediate card acquisition group whose urine samples at week 12 (end of the intervention) had detectable levels of THC, CBD, or metabolites with those of participants in the delayed card acquisition group without detectable levels of THC, CBD, or metabolites at week 12.

Results
Of

Sensitivity Analyses and Adverse Events
There were no significant factors that indicated which participants in the immediate card acquisition group would obtain a card (n = 120) vs which participants would not (n = 53) (eTable 9 in Supplement 2), and weighting the primary outcomes by propensity scores to calculate who would obtain a card yielded nearly identical estimates with no change in inferences (eTable 10 in Supplement 2).Comparing the primary outcomes in participants in the immediate card acquisition group with detectable levels of THC, CBD, or metabolites in urine at week 12 vs participants in the delayed card acquisition group without detectable levels of THC, CBD, or metabolites also yielded results that were similar to those presented in the main analyses (eTable 11 in Supplement 2).
Sensitivity analyses that included only assessments with confirmed blinded raters also yielded similar results (eTable 12 in Supplement 2).
One or more AEs were reported by 85 of 105 participants (80.1%) in the immediate card acquisition group and by 60 of 81 participants (74.1%) in the delayed card acquisition group Manual of Mental Disorders (Fifth Edition) (score range: 0-11, with higher scores indicating more severe CUD).This measure was analyzed in all participants.
c Pain severity was assessed with the severity subscale of the Brief Pain Inventory (score range: 0-10, with 10 being the worst imaginable pain).This measure was analyzed only in participants with a primary concern of pain.
d Insomnia symptoms were assessed with the Athens Insomnia Scale (score range: 0-24, with higher scores indicating more severe sleep difficulties).This measure was analyzed only in participants with a primary concern of insomnia.(eTables 13-14 in Supplement 2).A serious AE (cardiac event) occurred in 1 participant in the immediate card acquisition group.

Discussion
In this single-blind, pragmatic RCT involving people who sought cannabis products to improve insomnia, pain, or anxiety or depressive symptoms, participants in the immediate card acquisition group developed a greater number of CUD symptoms and had a higher incidence and greater severity of CUD diagnosis over the 12-week trial after obtaining a card compared with those in the delayed card acquisition group.Individuals with affective disorders have 3.9-fold (95% CI, 2.8-5.3)higher odds of meeting CUD diagnostic criteria, 35 and bidirectional associations between cannabis use and depression have been reported. 32,33These data suggest that a medical marijuana card may pose a high risk or may even be contraindicated for people with affective disorders.This finding is important to replicate because depression has been reported as the third most common reason that people seek a medical marijuana card. 3e finding of improved self-reported insomnia with a large effect size merits further study.The endocannabinoid system has been described as critical in regulation of the circadian sleep-wake cycle, 36,37 including maintenance and promotion of sleep. 38Although RCTs of cannabis for primary insomnia are lacking, small effects of cannabinoids on secondary sleep outcomes have been reported. 2ose with a primary insomnia concern were unlikely to develop CUD, suggesting a potential clinical utility of cannabinoids for insomnia.Thus, further study of the effect of cannabinoids on people with primary insomnia is warranted, using objective measures and self-assessments of sleep, along with analysis of CUD symptoms over a period longer than this 12-week trial.
Although pain, anxiety, and depressive symptoms were the common reasons cited for the use of cannabinoids, we detected no substantial benefit of a medical marijuana card for any of these outcomes.This null finding of medical marijuana card ownership on pain severity, pain interference, and pain catastrophizing is consistent with a recent systematic review and expert consensus recommendation against the use of any cannabinoids for chronic pain. 390][41] As such, mechanistic pain trials are warranted.We did observe an effect of medical marijuana card ownership on mental well-being and perceived stress that may be relevant to these health concerns.This finding deserves follow-up, and the potential risk of CUD should be weighed against the potential benefit to perceived stress and well-being.No AEs related to psychotic symptoms, mania, hypomania, or suicidal ideation or behavior were observed in the immediate card acquisition group during the trial.

Limitations
This study has several limitations.We assessed the risks and benefits of medical marijuana card ownership among participants who chose from a variety of cannabis products at self-titrated doses and who were aided by community physicians who recommended cannabis products for pain, insomnia, anxiety, or depressive symptoms.Therefore, we cannot comment on the pharmacological effects and risks or benefits of specific cannabinoids at specific doses for the health concerns we examined.Determination of the cannabinoid doses used is challenging in this circumstance. 26We followed expert consensus guidelines 43 by using timeline follow-back methods to document the frequency of use, rather than quantity or potency, of cannabis, supplemented with sensitive assays of urine cannabinoids. 25Because there was no placebo cannabis and because all participants sought cannabis as a potential therapy for their symptoms, the trial design created bias toward a treatment effect that was attributable to expectancy, strengthening our confidence in the null findings for pain, depression, or anxiety symptoms.The inclusion criteria were based on self-reported symptoms and as such may not be generalizable to those with a formal diagnosis of primary insomnia, an anxiety disorder, or major depressive disorder.The study sample was relatively homogeneous in race and ethnicity and educational attainment.There was a high dropout rate in the immediate card acquisition group between randomization and baseline, reflecting the cost and administrative burden of obtaining a medical marijuana card.Although retention after randomization was higher in the delayed card acquisition group, retention from baseline was high in both groups.

Conclusions
In this RCT, ownership of a medical marijuana card led to rapid onset and increased incidence and severity of CUD in some participants, particularly those with a chief concern of anxiety or depressive symptoms.The self-reported improvement in sleep quality warrants further study into the benefits of medical marijuana card ownership for insomnia and the risk of CUD.There were no observed benefits of obtaining a medical marijuana card for pain, anxiety, or depressive symptoms.

Figure 1 .
Figure 1.Study Flow Diagram outcome and assessed by our doctorate-level (J.M.G., R.M.S., and M.E.C.) or nursing-level (S.H.) study staff, who recommended CUD treatment referral and reduced cannabis use for participants who developed moderate or severe CUD.

e
Depression and anxiety symptoms were assessed with the Hospital Anxiety and Depression Scale (score range: 0-21, with 0-7 indicating normal, 8-10 indicating borderline abnormal [borderline anxiety or depression], and 11-21 indicating abnormal levels).This measure was analyzed only in participants with a primary concern of depression or anxiety.

Figure 3 .
Figure 3.Effect Sizes for Primary, Secondary, and Exploratory Outcomes

Table 1 .
Participant Characteristics by Randomization Group acquisition group who obtained a card and had 1 postbaseline assessment.Analyses used generalized estimating equations to account for repeated measures for a given individual across the primary, secondary, and exploratory outcomes.The statistical model assumed a constant effect of a medical 31rijuana card over time (weeks 2, 4, and 12 visits), with baseline scores as a covariate.A logistic regression model was used to estimate the odds ratio (OR) for CUD diagnosis, and linear models were used for continuous outcomes to estimate the mean difference (MD) in symptom scores.All tests and CIs were 2 sided, and statistical significance was defined as a P Յ .05 for the primary outcomes.We reported original30and adjusted P values that were corrected for multiple comparisons.31Results of secondary or exploratory analyses were reported as point estimates with 95% CIs.Analyses were performed using R, version 4.0.2(RFoundation for Statistical Computing).We conducted 2 post hoc sensitivity analyses.First, owing to a substantial number of dropouts in the immediate card acquisition group (participants who did not obtain a card; 53 of 173 [30.6%]), we ran a multivariable logistic regression to test whether baseline characteristics (eg, age, sex, race and ethnicity [which were self-reported and included African American or Black; Asian; Hispanic; multiracial; Pacific Islander; White; or unknown, including missing race and ethnicity information], As expected, the immediate card acquisition group reported significantly greater cannabis use throughout the intervention period than the delayed card acquisition group (likert scale difference: 2.44; 95% CI, 2.08-2.81;P<.001)(Figure2AandeFigure in Supplement 2).Participants in the immediate card acquisition group reported a greater number of CUD symptoms over the 12-week trial than those in the delayed card acquisition group (MD, 0.28; 95% CI, 0.15-0.40;P<.001)(Table2and Figure 3; eFigure in Supplement 2).Participants in the immediate card acquisition group had reduced self-rated insomnia symptoms over the 12-week intervention compared with those in the delayed card acquisition group (MD, -2.90; 95% CI, -4.31 to -1.51; P < .001).There was no significant group effect on pain, anxiety, or depressive symptom ratings (Table2 and Figure3).eTables3and4 in Supplement 2 show the self-reported methods of use and urine cannabinoid metabolite measurements in each group.The immediate card acquisition group had greater score improvement in mental well-being on the SF-12 (MD, 4.67; 95% CI, 2.63-6.71;Cohend= 0.39) but showed no significant effect on physical well-being compared with the delayed card acquisition group.There were no significant group effects on cognitive task performance (all with Cohen d Յ0.2) (eTable 5 in Supplement 2; Figure3).
and at least 1 postbaseline visits and were included in the analyses (immediate card acquisition group: n = 105; delayed card acquisition group: n = 81).Table 1 and eTables 1 and 2 in Supplement 2 provide the characteristics of the participants.Participants in the immediate card acquisition group were more likely to develop a Diagnostic and Statistical Manual of Mental Disorders (Fifth Edition)-based CUD diagnosis over the trial period vs those in the delayed card acquisition group (adjusted OR, 2.88; 95% CI, 1.17-7.07;P=.02).Throughout the 12 weeks, 18 participants (17.1%) in the immediate card acquisition group had a CUD diagnosis during at least 1 time point vs 7 participants (8.6%) in the delayed card acquisition group.Within the anxiety or depressive symptom subgroup, throughout the 12 weeks, 13 of 46 participants (28.3%) in the immediate card acquisition group and 4 of 37 participants (10.8%) in the delayed card acquisition group developed CUD (Figure2B).Most CUD diagnoses were mild (eTable 6 in Supplement 2).eTable 7 in Supplement 2 includes a count of the CUD symptoms that were endorsed within each group.The immediate card acquisition group reported greater CUD symptom severity on the CUDIT than the delayed card acquisition group, with a large effect size (CUDIT score MD, 2.22; 95% CI, 1.65-2.78;Cohend= 0.8) (eTable 8 in Supplement 2).The immediate card acquisition group had higher scores on the CGI-Improvement subscale, with a medium effect size (MD, 0.37; 95% CI, 0.56-0.19;Cohend = 0.6) and no notable difference from the delayed card acquisition group in scores on the CGI-Severity subscale.The immediate card Figure 2. Frequency of Cannabis Use and Incidence of Cannabis Use Disorder (CUD) Diagnoses in Immediate vs Delayed Card Acquisition Groups A Incidence of CUD by randomization group and primary complaintA, Cannabis use was assessed via a self-reported scale, which asked for frequency of cannabis use at each visit.There was a significant increase in use in the immediate card acquisition group vs the delayed card acquisition group (2.44; 95% CI, 2.08-2.81;P < .001).B, Cannabis use disorder was defined as 2 or more CUD symptoms on an 11-point scale.The odds of developing CUD were 2.9-fold higher in the immediate card acquisition group vs the delayed card acquisition group (adjusted odds ratio, 2.88; 95% CI, 1.17-7.07;P = .02).

Table 2 .
Primary Outcomes by Randomization Group b CUD symptoms were assessed with the CUD Checklist for Diagnostic and Statistical 4ffect of Medical Marijuana Card Ownership on Pain, Insomnia, and Affective Disorder Symptoms This study indicated that obtaining a medical marijuana card and using cannabis products from a dispensary, with the required medical oversight, for medical concerns of pain, anxiety, or depressive symptoms increased the risk for developing CUD without significantly improving symptoms.The odds of developing CUD were almost 2.9-fold higher in the immediate card acquisition group than in the delayed card acquisition group.Epidemiologic surveys of recreational cannabis use found that 3 in 10 adults who use cannabis develop CUD,4but it is unknown whether adults with a medical marijuana card would develop CUD at a lower rate than recreational users.In this trial, 17.1% of participants in the immediate card acquisition group developed a CUD diagnosis throughout the 12 weeks of study after acquiring a card.Although this incidence is lower than the 12-month incidence of CUD reported by Hasin et al,4the current study assessed the onset of CUD after only 12 weeks of medical marijuana card ownership.Thus, further research is warranted with a longer follow-up of people who obtain a card to better understand CUD risk in this group.This trial showed that CUD can develop at a fast rate within the first 12 weeks of medical marijuana card ownership, suggesting that those with a card develop CUD at similar rate as those who use cannabis recreationally and that the motive for use (eg, medical) may not be protective.Although most cases of CUD onset in the trial were mild, with 2 to 4 symptoms, these symptoms developed over a short, 12-week initial exposure.The most commonly reported CUD symptoms were higher tolerance and continued use despite the recurrent physical or psychological problems caused or exacerbated by cannabis.Most of those who developed CUD sought a medical marijuana card for affective symptoms;28.3% of participants in the immediate card acquisition group with a chief concern of anxiety or depression and 10.8% of participants in the delayed card acquisition group experienced CUD onset during the 12-week study.Thus, consistent with findings from epidemiologic studies, 32,33 people with affective symptoms who have access to cannabis through medical marijuana cards may be particularly at risk of CUD onset.Cannabis use disorder frequently co-occurs with affective disorders, particularly depression.
Cohen d and 95% CIs were obtained from generalized estimating equations linear models.For clarity, all outcomes are plotted on the x-axis by worsening or improving, rather than by item score.JAMA Network Open | PsychiatryJAMA Network Open.2022;5(3):e222106.doi:10.1001/jamanetworkopen.2022.2106(Reprinted) March 18, 2022 9/14 Downloaded From: https://jamanetwork.com/ on 09/23/2023 Effect of Medical Marijuana Card Ownership on Pain, Insomnia, and Affective Disorder SymptomsThe findings of this trial may provide insight into the potential risks and benefits of a medical marijuana card for people seeking cannabis for medical concerns.Although the development of CUD may suggest the likelihood of future medical marijuana card ownership, 42 participants in this study did not have evidence of problematic cannabis use at baseline, and yet some participants developed CUD over the first 12 weeks of card ownership.Thus, clinicians and patients are advised to consider the risks of cannabis use, especially in those with affective disorders, who may be particularly susceptible to developing CUD.In many US with marijuana laws, cannabis is approved for a variety of medical conditions, but little evidence of its efficacy is available.Thus, further study is needed to replicate and extend the findings of this trial, specifically the risks and benefits of cannabis use for specific medical conditions and the rate of CUD development over longer periods and under different health conditions, particularly mental health conditions.Such work can inform clinical decision-making on whether to initiate cannabis for specific medical concerns.
JAMA Network Open.2022;5(3):e222106.doi:10.1001/jamanetworkopen.2022.2106(Reprinted) March 18, 2022 10/14 Downloaded From: https://jamanetwork.com/ on 09/23/2023 Kim JH, Weinberger AH, Zhu J, Barrington-Trimis J, Wyka K, Goodwin RD.Impact of state-level cannabis legalization on poly use of alcohol and cannabis in the United States, 2004-2017.Drug Alcohol Depend.2021;218: 108364.doi:10.1016/j.drugalcdep.2020.10836443.Loflin MJE, Kiluk BD, Huestis MA, et al.The state of clinical outcome assessments for cannabis use disorder clinical trials: a review and research agenda.Drug Alcohol Depend.2020;212:107993.doi:10.1016/j.drugalcdep.Medical and Psychiatric History of the Overall Study Sample eTable 2. Concomitant Medications at Enrollment by Intervention Group eTable 3. Self-Reported CBD:THC Ratio and Method of Use by Intervention Group eTable 4. Cannabinoid Metabolite Measurements at Week 12 by Intervention Group eTable 5. Secondary Outcomes (SF-12 Mental and Physical Well-Being, and Cognitive Outcomes) by Intervention Group eTable 6. CUD Diagnosis by Group and Study Timepoint eTable 7. Cannabis Use Disorder Symptoms Endorsed in Each Group eTable 8. Exploratory Cannabis Use and Symptom Rating Outcomes eTable 9. Sample Characteristics at Screening Those Randomized to MMC Who Obtained vs Did Not Obtain MMC eTable 10.Sensitivity Analysis of Primary Outcomes by Randomization Group With Adjustment Using Propensity Scores eTable 11.Sensitivity Analysis of Primary Outcomes in MMC Participants With Metabolites, and in WLC Participants Without Detectable Metabolites eTable 12.Primary Outcomes in Participants With Blinded Assessments eTable 13.Treatment Emergent Adverse Events eTable 14.Severity of Treatment Emergent Adverse Events eFigure.CUD Symptoms by Intervention Group at Each Study Timepoint eReferences 42.