Mystical-type experiences occasioned by ketamine mediate its impact on at-risk drinking: Results from a randomized, controlled trial

Alcohol misuse imposes substantial medical, psychological and economic burdens, and current treatments for alcohol use disorder (AUD) — behavioural interventions and medications — have only modest effects. Rothberg and colleagues note that integrating pharmacological agents with behavioural therapies may improve outcomes. Prior work, including a recent randomised controlled trial by the study team, indicated that a single subanesthetic infusion of the NMDA antagonist ketamine can promote sustained abstinence when delivered alongside a five-week course of motivational enhancement therapy (MET). The mechanisms underlying durable benefits of ketamine remain uncertain; plausible neurobiological candidates include modulation of glutamate signalling, enhanced neuroplasticity, and alterations in default mode network connectivity, while psychoactive effects at subanesthetic doses include dissociation, psychotomimetic phenomena and so-called mystical-type experiences. This analysis tests whether mystical-type effects, as measured immediately after infusion, mediate the impact of ketamine on subsequent drinking behaviour in alcohol-dependent adults enrolled in the above trial. The investigators compared a single 52-minute IV infusion of ketamine (0.71 mg/kg) with an active control, midazolam (0.025 mg/kg), given on a pre-specified “quit day” during week 2 of MET. They predicted that mystical-type phenomena, rather than dissociative effects, would account for ketamine’s efficacy in reducing at-risk drinking.

This paper reports a secondary analysis of a five-week, randomised, double-blind, placebo-controlled trial conducted at the New York State Psychiatric Institute between September 2014 and September 2017. Forty adults seeking treatment for alcohol dependence (DSM-IV) were randomised 1:1 to receive a single IV infusion of either ketamine or the active control midazolam on day 2 of week 2; the infusion was administered on a participant-selected “quit day” while participants were engaged in twice-weekly MET visits. Participants abstained from alcohol for at least 24 hours before infusion and fasted overnight. Relaxation and mindfulness exercises were used to prepare participants, and staff blinded to treatment assignment administered infusions. Participants were informed they might receive one of several possible medications to obscure assignment and reduce expectancy effects. Eligible participants met minimum drinking criteria (at least four heavy drinking days in the prior seven days or weekly thresholds of 35 drinks for males and 28 for females) and were excluded for physiological dependence on another substance or for active major depressive disorder, bipolar disorder or psychotic disorders. Urine toxicology and ethyl glucuronide testing were used to confirm abstinence and assess drug use. Clinical monitoring occurred twice weekly (three visits during infusion week), and participants received phone follow-ups at 1 and 3 months. Psychoactive effects were assessed immediately post-infusion using the Hood Mysticism Scale (HMS), a 32-item instrument adapted to capture infusion-related mystical-type phenomena across eight subscales (ego quality, unifying quality, inner subjective quality, temporal/spatial quality, noetic quality, ineffability, positive affect and religious quality), and the Clinician Administered Dissociative States Scale (CADSS), a 27-item measure of dissociative phenomena. The primary clinical outcome for this analysis was a binary indicator of “at-risk” drinking during the 21-day post-infusion follow-up, defined by NIAAA thresholds (more than four drinks on any day or 14 per week for men, more than three drinks on any day or seven per week for women). Secondary outcomes included time to relapse, defined as days from infusion to the first heavy drinking day or dropout (a right-censored time-to-event variable ranging 1–21 days). Subscale scores from the HMS were examined in relation to drinking outcomes. Statistical analyses were performed in SPSS v24 using listwise deletion for missing data. Because CADSS showed positive skew, robust regression methods were used where appropriate. Group differences in HMS and CADSS were tested with one-way ANOVA using robust regression. The effect of treatment on at-risk drinking was assessed using binary logistic robust regression. Mediation was evaluated with multivariate logistic robust regression including treatment assignment, HMS and CADSS as predictors; mediation was inferred if the putative mediator remained significant while the treatment effect diminished. Time-to-relapse differences were examined with a Kaplan–Meier survival analysis and log-rank test; Cox proportional hazards models tested whether HMS or CADSS predicted relapse, with proportional hazards assumptions checked via time-covariate interactions. Spearman correlations explored relationships between the eight HMS subscales and drinking outcomes.

Forty participants were randomised and post-infusion drinking data were available for all; however, missing questionnaire data reduced the sample for some analyses (three participants missing HMS and two missing CADSS), leaving 35 individuals in the primary mediation analyses. Baseline average alcohol use across the sample was 6.6 drinks per day. No adverse events related to the infusions were reported. Compared with midazolam, ketamine produced significantly greater mystical-type and dissociative effects when measured immediately after infusion. The ketamine group scored on average 38.15 points higher on the HMS (ß=38.1490, p<0.05) and 12.955 points higher on the CADSS (ß=12.955, p<0.05). Ketamine also reduced the odds of engaging in at-risk drinking during the 21-day follow-up by approximately five times (p<0.05). On survival analysis, participants who received ketamine had a significantly longer time to relapse compared with midazolam (log-rank χ2=4.133, p=0.042). HMS and CADSS scores were moderately correlated (R=0.489, p<0.01). However, HMS scores correlated negatively with the number of heavy drinking days post-infusion (R=-0.466, p<0.05), while the correlation between CADSS and heavy drinking days did not reach significance (R=-0.315, p=0.054). In mediation analyses, the HMS score emerged as a significant mediator of the relationship between treatment assignment and at-risk drinking (Exp(B)=1.045, p<0.05); CADSS did not mediate this relationship (p=0.742). Consistent with mediation, the direct effect of treatment on at-risk drinking lost significance once HMS was included in the multivariate model (treatment p=0.719). Analyses of HMS subscales within the ketamine arm (n=17) identified specific associations with outcomes: ineffability scores correlated with percentage of heavy drinking days post-infusion (rs=0.624, p=0.01), while positive affect scores correlated with percentage of days abstinent (rs=0.613, p<0.05) and with average number of daily drinks post-infusion (rs=0.554, p<0.05).

Rothberg and colleagues interpret the findings as support for a role of mystical-type psychoactive effects in mediating ketamine’s beneficial impact on drinking in alcohol-dependent individuals engaged in MET. The data indicate that mystical-type experiences measured immediately after infusion, but not dissociative phenomena, accounted for reductions in at-risk drinking and longer time to relapse, suggesting that a specific subset of subjective effects may contribute to enduring behavioural change. These results extend prior laboratory findings in cocaine-dependent volunteers and parallel observations linking psilocybin-occasioned mystical experiences to therapeutic benefit in other populations. The authors propose several psychological pathways through which mystical-type experiences might aid recovery: enhanced motivation to change, a refreshed sense of self and values, greater connectedness to others and the world, and existential reappraisal that supports altered decision-making. They note that pairing ketamine with MET, a therapy designed to elicit “change talk” and goal-directed reflection, may provide a framework for translating infusion-occasioned experiences into concrete behavioural change. The ineffability dimension of mystical experience was highlighted as potentially especially relevant, given its correlation with heavy drinking metrics in the sub-analysis. Several limitations are acknowledged. Causality cannot be definitively established: shared neurobiological effects of ketamine might produce both mystical-type experiences and therapeutic benefit, making the mystical experience a marker rather than a causal mechanism. Mood enhancement from ketamine could also account for both increased mystical-type scores and improved drinking outcomes; although the trial screened out participants with DSM-IV depressive disorder to reduce confounding by antidepressant effects, subclinical mood changes may still have influenced results. Generalisability is limited by a small sample size and the selection of a non-depressed subset of the AUD population; many people with AUD have comorbid psychiatric disorders that were excluded here. Finally, the trial combined ketamine with MET, so interactions between the drug’s subjective effects and the psychotherapeutic platform might have been important and warrant further study. The investigators suggest future research should dissect existential and psychospiritual mechanisms, examine multiple facets of mystical experience in relation to outcomes, and explore how subjective effects of ketamine and other psychedelics interact with psychotherapy to influence clinical endpoints across substance use and mental health disorders.

This secondary analysis strengthens the evidence that mystical-type phenomena occasioned by ketamine are associated with subsequent reductions in problematic drinking when ketamine is delivered within a motivational enhancement therapy programme. The authors conclude that mystical-type experiences merit further investigation as potential psychological mechanisms of change in addiction treatment, and that future studies should more precisely characterise these subjective effects and their interaction with psychotherapeutic interventions across diverse clinical populations.