A sub-set of psychoactive effects may be critical to the behavioral impact of ketamine on cocaine use disorder: results from a randomized, controlled laboratory study

Ketamine has shown sustained effects in several psychiatric conditions when given at sub-anesthetic doses, including possible anti-addiction effects. However, its transient psychoactive effects have usually been viewed as unwanted, because they are thought to increase abuse liability and behavioural impairment. Earlier research had nonetheless suggested that some of these effects, particularly mystical-type experiences, might be related to therapeutic benefit. The paper also notes that prior work in cocaine-dependent volunteers hinted that ketamine may reduce craving partly through such experiences, but those findings were limited by small sample size, a narrow assessment of psychoactive effects, and a lack of behavioural outcomes. Dakwar and colleagues set out to revisit a completed laboratory trial in a larger sample of cocaine-dependent research volunteers to test whether mystical-type experience mediates ketamine’s effects on cocaine-related outcomes. They also aimed to examine dissociative effects and near-death experience-type phenomena as possible mediators, and to determine whether ketamine produces these different kinds of acute subjective effects relative to an active control, midazolam. The broader goal was to clarify whether a subset of ketamine’s experiential effects may be tied to its anti-addiction impact, rather than being merely incidental side-effects.

This was a randomised, controlled, double-blind crossover laboratory study in 20 non-depressed, cocaine-dependent individuals who were not seeking treatment or abstinence. Complete psychoactive-effect data were available for 18 participants. The participants were medically screened, admitted to a controlled inpatient research unit for up to three separate 6-day hospitalisations, and each admission was separated by 2 weeks to reduce carry-over effects and allow assessment of cocaine use in the natural environment. Each hospitalisation followed a structured sequence. There was an initial 2-day washout period, then a cocaine sampling session on Day 3 in which two 25 mg free-base cocaine doses were smoked to establish value and intensify craving. On Day 4, participants received a 52-minute intravenous infusion. On Day 5, they completed a cocaine choice session involving five opportunities to choose between 25 mg cocaine and $11. Day 6 was discharge. The first hospitalisation used a saline sham infusion to identify participants who did not robustly choose cocaine before active drug phases. During the second and third hospitalisations, participants were randomised in counter-balanced order to ketamine or midazolam under double-blind conditions. Ketamine was given as a 0.11 mg/kg 2-minute bolus followed by 0.60 mg/kg over 50 minutes; midazolam was given as a 2-minute saline bolus followed by 0.025 mg/kg over 50 minutes. No psychotherapy or behavioural treatment was provided. The main outcome was reduction in cocaine self-administration. Secondary outcomes were reduction in cocaine craving and cocaine use in the natural ecology, assessed during 2 weeks after each hospitalisation by thrice-weekly follow-up visits with urine toxicology and questionnaires. The authors also calculated a composite global improvement score, weighting self-administration most heavily, followed by naturalistic cocaine use and craving. Percent improvement was measured relative to baseline conditions established after the saline phase. Psychoactive effects were assessed 20 minutes after infusion using three validated scales: the Clinician Administered Dissociative States Scale (CADSS), the Hood Mystical Experiences Scale (HMS), and the Near-Death Experiences Scale (NDES). Statistical analyses used SAS. The authors first compared ketamine with midazolam on psychoactive scales and global improvement using paired t tests. They then performed a mediation analysis with participant as a random effect, entering dose and all psychoactive variables into a multivariate model with global improvement as the dependent variable. Mediation was considered supported if the proposed mediator remained significant in the full model and the intervention effect weakened or lost significance after adjustment.

Complete data on psychoactive effects and cumulative response were available for 18 participants, who were described as having high baseline cocaine use and otherwise being psychiatrically uncomplicated. All participants tolerated the procedures without adverse events, including unexpected psychiatric disturbances or evidence of ketamine or benzodiazepine misuse. Ketamine produced significantly greater acute psychoactive effects than midazolam on all three measures. Mean HMS scores were 103 (s.e. 7.8) with ketamine versus 63 (s.e. 9.8) with midazolam, NDES scores were 10.7 (s.e. 1.6) versus 4.3 (s.e. 0.8), and CADSS scores were 24.5 (s.e. 3.4) versus 4.9 (s.e. 1.2); all comparisons were p < .001. Ketamine also produced significantly greater global improvement in cocaine-related outcomes than midazolam, with a score of 56% (s.e. 7%) compared with 20.7% (s.e. 3.8%), p < .001. This composite reflected decreases in cocaine self-administration, cocaine use in the natural ecology, and craving. In the mediation analysis, HMS was the only significant mediator of the treatment effect (b = 0.431, p = .0175). Ketamine dose was no longer significant in the full model (p = .0562), which the authors took as supporting mediation. The total indirect effect was 0.2029, accounting for 35.7% of the total treatment effect. By contrast, the dissociation and near-death experience measures did not mediate the behavioural outcome. The paper also reports that the detailed in-analysis conclusion from the pre-registered or prior primary trial was that ketamine reduced cocaine self-administration, craving, and use in the natural environment more than the active control. In this secondary analysis, the key additional finding was that the mystical-type score, rather than dissociation or near-death experience phenomena, accounted for the observed relationship between ketamine and overall improvement.

The authors interpret the findings as evidence that mystical-type experience may play a central role in ketamine’s behavioural impact on cocaine use disorder. They argue that ketamine reliably elicited several transient psychoactive effects, but only the mystical-type component, as measured by the HMS, was linked to sustained improvement in cocaine-related outcomes. They also emphasise that this was shown using behavioural endpoints rather than self-report alone, which they view as stronger evidence than earlier work based only on mood or motivation measures. In positioning their results relative to previous research, the authors note that earlier studies had linked ketamine’s dissociative or mystical-type effects to antidepressant or antiaddiction outcomes, but were limited by small samples, abbreviated measures, or different routes and doses of administration. They describe the present study as stronger evidence that ketamine can occasion mystical-type phenomena and that these phenomena may be relevant to behaviour change. They also note that although near-death experience-type phenomena were produced, they did not mediate outcome, and they suggest this may reflect differences in what the respective questionnaires capture, especially because the HMS includes aspects of ontology, self-experience, and perspectival change that the NDES does not directly assess. A key limitation, according to the authors, is that the analysis cannot establish mechanism in a causal sense. They caution that mystical-type experience may be a sensitive marker of ketamine’s biological effect rather than the direct cause of improvement, and they mention possible shared neurobiological pathways such as prefrontal neurogenesis, default-mode network changes, and network synchrony. They also acknowledge that the sample was small and drawn from non-treatment-seeking cocaine-dependent volunteers in a highly controlled inpatient setting, which may limit generalisability. Even so, they argue the findings are consistent with a wider literature suggesting that medication-occasioned altered states can have lasting significance and may be therapeutically useful. The authors further state that the study adds to evidence that a single sub-anesthetic ketamine infusion can be administered safely under controlled conditions, with no observed behavioural toxicity or misuse in this sample. In terms of implications, the authors suggest that future medication development should not aim for compounds that are entirely free of psychoactive effects, but rather for ketamine-like agents with a more selective phenomenological profile that preserves potentially helpful mystical-type experiences while minimising abuse liability and behavioural toxicity. They also call for future studies to examine links between subjective effects and neural correlates, and to identify which specific mystical-type phenomena may matter most across different disorders.

The authors conclude that this is the first rigorous controlled study to show that medication-occasioned mystical-type phenomena may influence decision-making and behaviour in a sustained way, even without psychotherapy or behavioural treatment. They argue that future work should focus on ketamine-like medications that retain beneficial psychoactive properties while reducing harmful ones, and on clarifying how subjective experience, brain activity, and clinical benefit are related.