A single ketamine infusion combined with mindfulness-based behavioral modification to treat cocaine dependence: a randomized clinical trial

Cocaine use disorder remains a major public health problem with no FDA‑approved pharmacotherapies and limited success of standard behavioural treatments. Previous preclinical and human laboratory work has suggested that modulation of glutamate neurotransmission, and specifically N‑methyl‑D‑aspartate receptor (NMDAR) modulation, may affect cocaine‑related learning and reinforcement. Ketamine, a dissociative anaesthetic given at subanesthetic doses, has shown robust antidepressant effects and in laboratory studies in people with cocaine dependence it has been reported to increase motivation to quit, reduce craving, and decrease cocaine use over short periods. However, those benefits have typically been transient, and it has been proposed that combining ketamine with a behavioural intervention might extend and translate its effects into clinically meaningful, sustained change. Dakwar and colleagues designed this randomised clinical trial to test whether a single subanesthetic intravenous ketamine infusion (0.5 mg/kg over 40 minutes), given to adults with cocaine dependence initiating mindfulness‑based relapse prevention (MBRP), would improve clinically relevant outcomes compared with an active control (midazolam). The primary endpoints were end‑of‑study abstinence (two weeks confirmed by urine toxicology) and time to relapse (first use or dropout); secondary aims included effects on weekly cocaine use, craving, and safety/tolerability. The trial therefore evaluated ketamine as an adjunct to a manualised mindfulness‑oriented behavioural platform rather than as a stand‑alone medication.

This was a randomised, double‑blind, parallel‑group trial conducted at the New York State Psychiatric Institute between September 2011 and December 2016. Fifty‑five treatment‑seeking adults who met DSM‑IV criteria for cocaine dependence and met prespecified recent‑use thresholds were enrolled. Key exclusions included current DSM‑IV depressive or anxiety disorders, psychotic or dissociative history, and benzodiazepine or opioid use disorder; participants were required to be medically healthy and under age 70. Screening included structured diagnostic interview, medical assessment, serum tests, and scales for depression and dissociation. Participants were hospitalised for a 5‑day inpatient phase during which they initiated an adapted 5‑week course of MBRP. On day 2 they received a single 40‑minute intravenous infusion of either ketamine hydrochloride 0.5 mg/kg or active control midazolam 0.025 mg/kg; infusions were prepared and administered by blinded staff. To preserve blinding and minimise expectancy effects, participants were told they might receive one of several possible medications. Vital signs were continuously monitored, and dissociative symptoms were assessed immediately postinfusion using the Clinician‑Administered Dissociative States Scale (CADSS). A morning MBRP session occurred about 2 hours postinfusion, with further high‑density MBRP sessions on days 2–5 and outpatient MBRP once weekly during weeks 2–5; participants returned twice weekly during weeks 2–5 for MBRP, physician visits and measures, and urine toxicology. A telephone follow‑up interview occurred at 6 months. The prespecified primary outcome was two weeks of end‑of‑study abstinence confirmed by urine toxicology. Time to relapse (first use or dropout after the inpatient phase) was analysed as a secondary primary outcome. Weekly cocaine use and weekly craving were analysed as secondary outcomes. Statistical methods included logistic regression for the primary abstinence outcome, Cox proportional hazards models for time to relapse (hazard ratios reported; a hazard ratio below 1 indicates reduced risk in the ketamine group), and longitudinal mixed‑effects models for weekly binary cocaine use and continuous craving scores (random intercepts; route of use included as a covariate). The investigators planned for a sample of 60 but stopped enrollment at 55 after an interim analysis indicated a robust difference that would remain significant under a conservative algorithm. Analyses were two‑sided with α=0.05 and were performed in SAS 9.4.

Fifty‑five participants were randomised (ketamine N=27; midazolam N=28). Baseline median daily cocaine expenditure was similar between groups (ketamine $32.86; midazolam $36.43). The proportion of freebase (smoked) users differed modestly between groups (55.6% in ketamine; 71.4% in midazolam). The infusion produced greater acute dissociative effects in the ketamine arm: median CADSS score postinfusion was 22 (IQR=13–34) versus 7 (IQR=4–10) for midazolam (p<0.001); psychoactive effects resolved within 30 minutes, and there was no persistent dissociation on postinfusion measures. On the primary abstinence outcome (intent‑to‑treat), 48.2% of participants in the ketamine group maintained abstinence over the last two weeks of the trial compared with 10.7% in the midazolam group. Time‑to‑event analysis showed the ketamine group had a reduced risk of relapse (first use or dropout) with a hazard ratio of 0.47 (95% CI=0.24, 0.92), indicating a 53% lower likelihood of relapse compared with midazolam. In longitudinal analyses controlling for route of use, odds of cocaine use in the midazolam group were 7.8 times those in the ketamine group (odds ratio=7.8; 95% CI=1.5, 39.9; p=0.01). There was no significant time-by‑treatment interaction for cocaine use, and no overall change in use over study weeks, consistent with maintenance of early gains in the ketamine arm. Craving scores were substantially lower in the ketamine group across the trial: the main‑effects model estimated craving to be 58.1% lower in ketamine than in midazolam (95% CI=18.6, 78.6; p=0.01), with no significant week‑by‑treatment interaction. Measures of perceived stress and mindfulness did not differ between groups over time. At the 6‑month telephone follow‑up, 44% (12/27) of ketamine recipients reported abstinence, versus 0% in the midazolam group (χ2=15.92, p<0.001). Infusions were reported to be well tolerated and no participants were withdrawn due to adverse events.

Dakwar and colleagues report that a single subanesthetic ketamine infusion, administered within a structured mindfulness‑based relapse prevention programme, was well tolerated and associated with improved cocaine‑related outcomes relative to an active control. The ketamine arm showed higher end‑of‑study abstinence rates, reduced risk of relapse (hazard ratio=0.47), markedly lower craving throughout follow‑up, greater retention and a higher self‑reported abstinence rate at 6 months. The investigators interpret these results as consistent with the hypothesis that ketamine can temporarily reduce craving, increase motivation, and decrease behavioural reactivity, and that embedding the drug within a behavioural framework such as MBRP may leverage these transient effects into more durable clinical gains. The authors situate their findings against prior laboratory studies that found short‑lived benefits of ketamine in cocaine‑dependent volunteers and argue that combining pharmacological and psychological interventions may extend benefits. They note possible mechanistic pathways relevant to both ketamine and mindfulness training (for example, modulation of mesolimbic function, promotion of prefrontal plasticity and synaptogenesis, and changes in default‑mode network connectivity). Potential biological mediators discussed include brain‑derived neurotrophic factor and, preliminarily, opioid‑related mechanisms; the investigators also raise the issue of ketamine’s abuse liability and describe safeguards used in the trial (careful patient selection, monitoring and postintervention support). Key limitations acknowledged by the authors include the relatively small sample size and some baseline imbalances (including route of use), the highly monitored inpatient initiation phase that limits generalisability to usual outpatient care, and the homogeneous sample composition (majority male and African American with minimal psychiatric comorbidity). They also note that the trial did not include a medication‑only arm, so it is not possible to determine whether MBRP was necessary to sustain benefits, nor was the integrity of the blind formally tested; informing participants that they might receive several different medications constituted deliberate deception to protect blinding but would not be practicable in routine clinical settings. The authors conclude that the results are promising but require replication in larger, more diverse samples, with study designs able to test the hypothesised synergy between ketamine and behavioural treatments and to clarify mechanisms and risks.