This randomised quadruple-blind placebo-controlled clinical trial (n=40) found that a single 25 mg/70 kg dose of psilocybin, given with psychotherapy, increased cocaine-free days and delayed relapse compared with diphenhydramine in people with cocaine use disorder.
Papers cited by this study that are also in Blossom
Goodwin, G. M., Aaronson, S. T., Alvarez, O. et al. · New England Journal of Medicine (2022)
Importance
Cocaine use disorder is a serious public health problem and no medications have been proven effective for its treatment.
Objective
To evaluate psilocybin in the treatment of cocaine use disorder. It was hypothesized that psilocybin, compared with placebo, would yield a higher percentage of cocaine abstinent days, a greater likelihood of complete abstinence from cocaine, and a greater latency to first cocaine lapse through 180 days after end of treatment.
Design, setting, and participants
Randomized, quadruple-blind, placebo-controlled clinical trial at a major medical research center in the Deep South of the US. Participants were individuals with cocaine use disorder who were motivated to quit and without significant comorbidities, recruited between May 2015 and August 2023 with data collection completed in May 2024.
Interventions
Participants were randomized (1:1) to receive a single oral dose of psilocybin (25 mg per 70 kg of body weight) or active placebo (100 mg diphenhydramine). All participants received manualized psychotherapy that incorporated cognitive-behavioral treatment approximately 1 month before and 1 month after an all-day investigational drug treatment session.
Main outcomes and measures: Percentage of cocaine abstinent days, rates of complete cocaine abstinence, and time to first cocaine lapse through 180 days after end of treatment, assessed by timeline followback interview and confirmed with urinalysis. Hypotheses were formulated before data collection and analyses followed intention-to-treat principles.
Results
Of the 40 participants, 33 (82.5%) were men, the median (IQR) age was 50.0 (43.8-56.0) years, 33 (82.5%) were Black, and 7 (17.5%) were White. Most participants had lower socioeconomic status, with 26 participants (65%) having an annual income of $20 000 or less. Four participants were lost to follow-up, resulting in 36 participants who completed assessments through 180 days after end of treatment. Psilocybin recipients had a higher percentage of cocaine abstinent days (β = 28.95; 95% CI, 18.22-39.67; P < .001), greater likelihood of complete cocaine abstinence (odds ratio, 18.37; 95% CI, 1.92-2468.17; P = .007), and a reduced risk of cocaine lapse over time (hazard ratio, 0.28; 95% CI, 0.13-0.60; P = .001) than active placebo recipients. No serious adverse events occurred.
Conclusions and relevance
In this randomized clinical trial, psilocybin appeared to be safe and efficacious for treating cocaine use disorder among individuals from underrepresented and vulnerable populations. Further research is warranted to replicate and expand these findings.
Cocaine use disorder remains a major public health problem, with global cocaine use at record levels and no medications yet proven effective. The paper notes that while behavioural interventions such as contingency management can help, pharmacological options for stimulant use disorders are still lacking. It also places psilocybin in the context of earlier psychedelic research, including trials in alcohol use disorder and smoking cessation that suggested possible anti-addictive effects, although findings have not been uniform across studies. Hendricks and colleagues set out to test whether psilocybin, combined with manualised psychotherapy, could improve cocaine abstinence outcomes compared with an active placebo plus the same psychotherapy. They hypothesised that psilocybin would increase the proportion of cocaine-abstinent days, raise the likelihood of complete abstinence, and delay the first cocaine lapse through 180 days after treatment. The study is presented as, to the authors' knowledge, the first randomized trial of a psychedelic in cocaine use disorder, with particular emphasis on whether such treatment could be implemented in underrepresented and socioeconomically disadvantaged participants.
Hendricks and colleagues conducted a randomised, quadruple-blind, active placebo-controlled, parallel-group trial at the University of Alabama at Birmingham. Participants were recruited between May 2015 and August 2023, with data collection completed in May 2024. Eligibility focused on adults with cocaine use disorder who were motivated to quit and without significant comorbidities. Reported inclusion criteria included a Severity of Dependence Scale score of at least 3, no hallucinogen use in the past 3 years, and good general health. Exclusions included pregnancy or breastfeeding, current psychiatric disorders other than substance abuse or nicotine dependence, current hypertension, use of certain psychotropic medications, and personal or family history of psychotic or bipolar disorders. Participants were randomised 1:1 to a single oral dose of psilocybin 25 mg per 70 kg body weight or 100 mg diphenhydramine as an active placebo. The median psilocybin dose was 32 mg, with a range of 21.6-61.6 mg. All participants received manualised psychotherapy incorporating a client-centred approach used in prior psilocybin studies plus cognitive-behavioural treatment for cocaine use disorder. This included 4 to 5 preparation sessions of about 2 hours each, one all-day drug session after 7 days of cocaine abstinence verified by urinalysis, and 5 post-randomisation integration sessions of about 1 hour each, with follow-up assessments at 90 and 180 days after the final integration session. Outcome measurement relied on timeline followback interviews and urine toxicology. The primary outcomes were urinalysis-verified percentage of cocaine-abstinent days across specified periods, complete cocaine abstinence from the drug session through day 180, and time to first cocaine lapse. Cocaine abstinence was biochemically checked with urine benzoylecgonine testing, and adverse events were monitored throughout the study. Analyses followed intention-to-treat principles. The authors used mixed models for repeated measures for abstinent days, Fisher exact tests and Firth penalised logistic regression for complete abstinence, and Kaplan-Meier curves with log-rank testing and Cox proportional hazards modelling for time to lapse. No power calculation was conducted before the trial began.
This study was conducted at the University of Alabama at Birmingham (UAB), a major medical research center in the Deep South of the US, using a randomized, quadruple-blind, active placebocontrolled, parallel-group design. It was approved by the UAB institutional review board and participants provided written informed consent. The trial protocol and statistical analysis plan are available in Supplement 1 and Supplement 2, respectively. No patients or members of the public were involved in the study design or conduct. This report follows the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline for randomized trials.
Participants were recruited through media advertisements describing the study and seeking individuals who "use cocaine and want to quit." Chain referral was used to recruit 2 participants. Participants self-reported sociodemographic characteristics including gender, race, and ethnicity. by a goal of complete abstinence on the Thoughts about Abstinence Questionnaire.Additional inclusion criteria included a score of at least 3 on the Severity of Dependence Scale,no hallucinogen use in the past 3 years, and good general health as assessed by medical history interview and physical examination. Exclusion criteria included pregnancy or breastfeeding, current psychiatric disorders other than substance abuse or nicotine dependence, current hypertension, current use of certain psychotropic medications (eg, antidepressants, antipsychotics, and mood stabilizers), and personal or family history of psychotic or bipolar disorders. Full inclusion and exclusion criteria are provided in the trial protocol (Supplement 1).
Participants were randomized in a 1:1 ratio by UAB's Investigational Drug Service within 48 hours of the all-day session to receive either a single oral dose (25 mg per 70 kg of body weight) of psilocybin (median [range] dose, 32 [21.6-61.6] mg) or 100 mg diphenhydramine using a random number table. Allocation concealment was maintained without further study involvement by the UAB Investigational Drug Service. Participants, therapists, outcome assessors, and independent statisticians were masked to group assignment. Study medications were administered in identical opaque capsules. Participant expectancies for the treatment they were receiving were assessed with the Credibility/Expectancy Questionnaire (CEQ) after the final preparation session, before the first postrandomization integration session, and after the final integration session. Blinding integrity was assessed by asking participants to guess whether they received psilocybin or diphenhydramine and to rate how confident they were in their guess (0, not at all; 1, a little; 2, moderate; or 3, very much) before the first integration session; both the primary and secondary therapists were similarly asked to guess whether participants received psilocybin or diphenhydramine then rate how confident they were in their guess after the final integration session.
Create a free account to open full-text PDFs.
Raison, C. L., Sanacora, G., Woolley, J. D. et al. · JAMA (2023)
Griffiths, R. R., Johnson, M. W. · Journal of Psychopharmacology (2016)
Ross, S., Bossis, A. P., Guss, J. et al. · Journal of Psychopharmacology (2016)
Anderson, B. T., Danforth, A. L., Daroff, R. et al. · EClinicalMedicine (2020)
Krebs, T. S., Johansen, P. Ø. · Journal of Psychopharmacology (2012)
Luquiens, A., Belahda, D., Graux, C. et al. · Addiction (2025)
Rieser, N. M., Bitar, R., Halm, S. et al. · EClinicalMedicine (2025)
Johnson, M. W., Naudé, G. P., Hendricks, P. S. et al. · JAMA Network Open (2026)
Ortiz, C. E., Dourron, H. M., Sweat, N. W. et al. · Neuropharmacology (2022)
Szigeti, B., Weiss, B., Rosas, F. E. et al. · Psychological Medicine (2024)
Nayak, S., Bradley, M. K., Kleykamp, B. A. et al. · Journal of Clinical Psychiatry (2023)
Forty participants were randomised, and 36 completed assessments through 180 days after the end of treatment. The sample was predominantly male (33/40, 82.5%), with a median age of 50 years, and was largely Black (33/40, 82.5%); 65% reported annual income of $20,000 or less. Four participants were lost to follow-up. The extracted text indicates that there was no baseline difference between groups in cocaine-abstinent days during the prerandomisation periods, supporting comparability before treatment. On the primary efficacy outcomes, psilocybin was associated with better cocaine outcomes than active placebo. The psilocybin group had a higher percentage of cocaine-abstinent days overall, with a reported β of 28.95 (95% CI, 18.22-39.67; P < .001). For complete abstinence from the integration period through day 180, 6 of 20 psilocybin participants (30%) versus 0 of 20 placebo participants achieved complete abstinence. Firth penalised logistic regression estimated that psilocybin recipients were about 18 times more likely to report complete abstinence (odds ratio, 18.37; 95% CI, 1.92-2468.17; P = .007), with a number needed to treat of 3.33. The authors note that the confidence interval was very wide because of the small sample and separation in the data. Time to first cocaine lapse was also longer with psilocybin: the hazard ratio was 0.28 (95% CI, 0.13-0.60; P = .001), indicating reduced lapse risk over time. The paper reports that no participants were completely abstinent during the prescreening period, and that during the preparation period 2 of 20 psilocybin participants and 3 of 20 placebo participants reported complete abstinence, but these differences were not significant. The extracted text also notes that there was no indication that participants substituted other drugs for cocaine, although those analyses were said to be reported separately. Four participants were lost to follow-up and were treated as nonabstinent for the complete abstinence outcome. No serious treatment-related adverse events occurred. Adverse events were more common in the psilocybin group than in placebo (13/20, 65% versus 2/20, 10%), but most were expected, occurred during the all-day session, and resolved without sequelae.
The authors interpret the findings as evidence that psilocybin, when paired with psychotherapy, may be safe and efficacious for cocaine use disorder. They emphasise that this was, to their knowledge, the first randomized clinical trial to show improvements in cocaine-abstinence days, complete abstinence, and time to lapse with psilocybin. They frame the results as an important advance for a condition with no approved pharmacotherapies and limited psychosocial treatment options. Hendricks and colleagues also highlight the significance of the participant sample. They state that psychedelic trials in the US have often included people with relatively high socioeconomic status, whereas this study involved Black participants and people with lower incomes who are disproportionately affected by cocaine use disorder. They present the trial as evidence that this type of treatment can be feasibly delivered in a more vulnerable and underrepresented population. The authors acknowledge several limitations. They note that the salient effects of psilocybin made blinding difficult, and it is unclear whether unblinding influenced outcomes. They also point out that the lead author served as the primary therapist, which could have introduced allegiance bias despite the use of manualised therapy. Reliance on self-reported cocaine use is another limitation, although urinalysis partly addressed this. They state that fidelity monitoring of psychotherapy was not conducted, so observed effects cannot be attributed solely to psilocybin rather than therapeutic factors. Generalisability may be limited because trial participants are a selected group, and the small sample produced wide confidence intervals, so the findings should be treated as hypothesis-generating rather than confirmatory. The authors call for adequately powered confirmatory trials, larger and more diverse samples, work on optimal dosing and therapeutic context, better psychotherapy fidelity monitoring, and pragmatic studies in real-world clinical settings.
The authors conclude that psilocybin appeared safe and efficacious for cocaine use disorder in this trial, but that these are preliminary findings. They state that larger confirmatory studies are needed to validate the results and to clarify how best to deliver the treatment in clinical practice.
Following telephone screening, participants completed 2 prescreening visits. Informed consent was obtained at the first visit, after which participants completed sociodemographic measures and an interview-administered retrospective calendar-based assessment using the timeline followback (TLFB) technique that queried cocaine and other drug use. Cocaine use was biochemically verified with urine samples testing for the presence of benzoylecgonine 300 ng/mL or greater, with a detection window of 24 to 240 hours after use.The same retrospective calendar-based assessment using the TLFB technique was interview-administered at all subsequent appointments and queried cocaine and other drug use since the last study visit, with collection of urine samples to biochemically verify cocaine abstinence at all study visits. Following completion of the 2 prescreening visits, participants completed a medical visit that included physical examination by a physician and a structured diagnostic interview for DSM-IV disorders 30 by trained staff. Qualified participants then completed 4 to 5 preparation psychotherapy sessions of approximately 2 hours each scheduled to occur once weekly, 1 all-day drug session preceded by 7 days of cocaine abstinence verified by urinalysis, 5 postrandomization integration psychotherapy sessions of approximately 1 hour each scheduled to occur once weekly and with the first integration session occurring within 48 hours after the all-day drug session, and follow-up assessments 90 days (day 90) and 180 days (day 180) after the final integration session. Psychotherapy was manualized by P.S.H. and integrated a client-centered approach used in prior psilocybin trialswith a cognitive-behavioral treatment for CUD.
The first primary outcome was urinalysis-verified percentage of cocaine abstinent days measured across 6 time periods: in the past 90 days at baseline; during the prescreening period (ie, baseline through the first preparation session); during the preparation period; during the integration period; from the final integration session through day 90; and from day 90 through day 180. The second primary outcome was complete cocaine abstinence, a dichotomous outcome defined as no use of cocaine whatsoever after the all-day drug session through day 180 verified by urinalysis and with participants lost to follow-up coded as nonabstinent. The final primary outcome was time to cocaine lapse, measured as days until first reported cocaine use since the all-day drug session. Participants were observed throughout the all-day drug session and specifically queried for any drug-related adverse events (AEs) at the outset of the first integration session. Any new or worsening AEs observed or reported during or after the all-day drug session were deemed related to the intervention based on temporal proximity, participant attribution, and investigators' judgment; AEs were deemed expected from the intervention based on literature review and investigators' judgment.
No power calculations were conducted before study initiation. As determined a priori, groups were tested on the first 2 primary outcomes before randomization to confirm group comparability and detect potential sources of bias, ensuring valid causal inference.No multiplicity adjustments were applied, consistent with recommendations that they can lead to more interpretive errors than transparent reporting without adjustment.Unadjusted 2-sided P values of .05 or lower were considered statistically significant. P values are reported to the nearest hundredth unless they are less than .01, in which case they are reported to the nearest thousandth, allowing for readers to apply their preferred multiplicity corrections. Mixed models for repeated measures (MMRMs) using first-order autoregressive covariance structures compared groups on percentage of cocaine abstinent days while accounting for longitudinal trends, within-participant variability, and missing data.An MMRM first compared groups across prerandomization time periods. An MMRM then compared groups across postrandomization time periods. This model was fitted with a group-by-time period interaction term for pairwise comparisons at each postrandomization time period. Estimated marginal means were extracted from this model, from which contrasts and standardized effect sizes (Hedges g) were calculated. Fisher exact test and Firth penalized logistic regressionfirst compared groups on complete cocaine abstinence from the prescreening period through the preparation period. Fisher exact test and Firth penalized logistic regression were then used to compare groups on complete cocaine abstinence from the integration period through day 180. Kaplan-Meier survival curves were plotted to show time to cocaine lapse by group, with a log-rank test performed for initial univariate comparison. A Cox proportional hazards model was then used to calculate a hazard ratio, assuming noninformative censoring. Analyses included all randomized participants per intention-to-treat and were conducted using R version 2024.04.02 (R Project for Statistical Computing). This trial was monitored by an independent data monitoring committee and prospectively registered at ClinicalTrials.gov, where it was first posted on January 15, 2014, before the first participant was enrolled. psychotherapy, only 15 of 20 placebo participants (75%) did the same. Of the 5 placebo participants who did not complete integration psychotherapy, 1 attended no integration sessions, 1 completed 1 integration session, 1 completed 2 integration sessions, and 2 completed 4 integration sessions.
Urine samples were labeled discordant if positive for the presence of benzoylecgonine 300 ng/mL or more despite no self-reported cocaine use in the past 7 days. Three participants yielded discordant urine samples: 1 psilocybin participant at second prescreening visit; 1 placebo participant at the second prescreening visit; and 1 placebo participant at the second prescreening visit and fourth preparation session. In each instance, participants were reminded that cocaine use would not be penalized and the priority of the study was to obtain accurate data. They were given the opportunity to modify their self-report but chose not to do so, and their data were entered as reported. None these participants reported complete abstinence in the prescreening or preparation periods and no further discordant urine samples were obtained.
Figurepresents the mean percentage of cocaine abstinent days across the study by group. An MMRM comparing groups across prerandomization time periods did not reveal a main effect of group (β = 0.94; 95% CI,- participants during the integration period, from the final integration session through day 90 and from day 90 through day 180, respectively. There was no indication that psilocybin participants substituted methamphetamine, nonmedical stimulants, or other drugs for cocaine (these results will be reported in a separate publication). No participants reported complete abstinence during the prescreening period, whereas 2 of 20 psilocybin participants (10%) and 3 of 20 placebo participants (15%) reported complete abstinence during the preparation period; Fisher exact test (P > .99) and Firth penalized logistic regression (odds ratio [OR], 0.67; 95% CI, 0.10-3.91; P = .65) failed to reveal an association between group and complete abstinence during this time. None of these 5 participants reported complete abstinence from the integration period through day 180. Six of 20 psilocybin participants (30%) and 0 of 20 placebo participants reported complete abstinence from the integration period through day 180; Fisher exact test (P = .02) showed a significant association between group and complete abstinence, with Firth penalized logistic regression indicating that psilocybin participants were approximately 18 times more likely to report complete abstinence than placebo participants (OR, 18.37; 95% CI, 1.92-2468.17; P = .007), and a number needed to treat of 3.33 (95% CI, 2.0-10.1). The wide CIs reflect the small sample size and data separation in this variable-outcome combination. Kaplan-Meier estimates (Figure) demonstrated survival probabilities of 55.0% (95% CI, 36.99%-81.75%) and 21.05% (95% CI, 8.81%-50.28%) at 90 days in the psilocybin and placebo a Adverse events were classified and reported using version 28.0 of the Medical Dictionary for Regulatory Activities, with preferred terms applied for event categorization. All-day drug session adverse events were recorded from drug administration to participant discharge. Hypertension was defined as 160 mmHg or more systolic pressure or 100 mmHg or more diastolic pressure at any assessment, and tachycardia was defined as 100 or more beats per minute at any assessment. Although post-drug session adverse events were recorded from all-day drug session discharge to the completion of study participation, all but suicidal ideation occurred within 24 hours after discharge. With regard to suicidal ideation, more than 1 month after the all-day drug session, this participant reported a lapse to cocaine use following a prolonged period of abstinence coupled with feelings of disappointment, frustration, and passive suicidal ideation, which was resolved with no sequelae. All other adverse events were similarly resolved with no sequelae. Although adverse events were monitored throughout the study, only potentially serious or drug-related events were systematically documented, which likely resulted in underreporting of mild and/or unrelated adverse events compared with the comprehensive safety data collection typical of registrational trials. groups, respectively (log-rank χ 2 1 = 11.65; P < .001). Participants in the psilocybin group had a significantly reduced risk of lapse compared with participants in the placebo group (hazard ratio, 0.28; 95% CI, 0.13-0.60; P = .001). AEs occurred in 13 of 20 psilocybin participants (65%) and 2 of 20 placebo participants (10%) (Table). No serious treatment-related AEs occurred. Most AEs were expected and occurred during the all-day drug session, and all were resolved with no sequelae. No rescue medications were administered.
This randomized clinical trial is the first, to our knowledge, to demonstrate that psilocybin coupled with psychotherapy may be safe and efficacious in the treatment of CUD. Psilocybin-treated participants showed significantly greater percentages of cocaine abstinent days, higher rates of complete abstinence from cocaine, and a decreased risk of cocaine lapse over time. These findings are a potentially important advancement in the treatment of CUD, a condition for which there are no approved pharmacotherapies and limited psychosocial interventions.The representation of vulnerable populations in psychedelic clinical trials has been a crucial ongoing concern.A recent systematic review found that participants in US-based psychedelic trials typically had higher socioeconomic status than the general population.The present study demonstrates that psilocybin treatment can be feasibly implemented with Black and socioeconomically disadvantaged individuals vulnerable to the adverse impacts of CUD but understudied in psychedelic research.
Strengths of this study include its randomized and blind approach, representative sample, active placebo comparator, high retention rates, urine toxicology to verify self-report, and intention-totreat analysis. However, salient drug effects presented challenges to blinding, an oft-discussed topic in psychedelic research.Whether there is a relationship between unblinding and efficacy is unclear.Additional limitations included the involvement of the lead author as the primary therapist, which may have introduced allegiance bias mitigated by a manualized therapy protocol,and reliance on self-reported cocaine use, which, though supplemented by confirmatory urinalysis, could have affected the accuracy of outcomes. While self-report generally shows high concordance with biological measures in research settings,including in 2 recent trials of psilocybin for AUD,underreporting remains possible. Fidelity monitoring of psychotherapy was beyond the scope of the current study, hindering attribution of observed effects to psilocybin rather than psychotherapeutic factors. Furthermore, as participants in clinical trials often represent a highly selected subset of patients, generalizability to routine clinical populations may be limited.Given the small sample size and resulting wide CIs around the effect size estimates, findings should be interpreted with caution and conceptualized as hypothesis-generating rather than confirmatory.
In this randomized clinical trial, psilocybin appeared safe and efficacious for CUD, a difficult-to-treat condition with significant personal and public health impacts. Adequately powered confirmatory trials are needed to validate these preliminary findings. Future research should employ larger samples across diverse populations, explore the optimal therapeutic context and dosing regimens to enhance outcomes, monitor the fidelity of psychotherapy to advance standardization, replicability, and understanding of the therapeutic process, and consider pragmatic trials to interrogate effectiveness in clinical settings.
Psilocybin in the Treatment of Cocaine Use Disorder Drug Administration authorization, and certain associated intellectual property has been exclusively licensed to Filament Health.