Human psychopharmacology and dose-effects of salvinorin A, a kappa opioid agonist hallucinogen present in the plant Salvia divinorum

Salvia divinorum has a long ethnomedical history among Mazatec shamans and its principal psychoactive constituent, salvinorin A, is a neoclerodane diterpene that acts as a potent, selective non-nitrogenous kappa opioid receptor agonist. Unlike classic serotonergic hallucinogens (which act primarily at 5-HT2A receptors), salvinorin A shows no 5-HT2A activity; animal studies report a kappa-like profile on antinociception and behaviour but have produced mixed evidence regarding abuse liability. Concurrently, recreational use of S. divinorum has increased in the United States and Europe, legal controls have expanded, and human data on the psychopharmacology of isolated salvinorin A have been limited. Johnson and colleagues set out to characterise the acute physiological, behavioural and subjective effects of inhaled salvinorin A across a wide ascending dose range in psychologically and physically healthy adults with prior hallucinogen experience. The study aimed to describe time course and dose–response relationships, assess safety and tolerability under supportive conditions, and compare subjective effects with those of classic hallucinogens, including measures of mystical-type experience.

Recruitment and screening took place in the Baltimore area. After a telephone pre-screen, volunteers underwent an in-person structured clinical interview administered by a social worker with psychiatric consultation available; exclusion criteria included current or past schizophrenia or psychotic disorder, bipolar I/II disorder, recent substance dependence (past 5 years, excluding caffeine and nicotine), severe major depression, dissociative disorder, and first- or second-degree relatives with psychotic or bipolar disorders. Medical screening comprised a physical exam, ECG, and blood tests. Four participants (mean age 29.5 years, range 23–35, two female) who reported prior S. divinorum use and had at least a high-school education completed all sessions. The extracted text does not clearly report participant bodyweights in full detail. Salvinorin A was isolated and reported as 99.5% pure by HPLC. Because very small amounts were required, measured quantities of salvinorin A were dissolved in acetone, filtered, and 1 ml of solution was evaporated inside a 5 ml round-bottom flask so that a residue remained on the flask interior. For administration, the flask was attached to a vacuum adapter with tubing through which participants inhaled; tubing was replaced between participants but not between sessions for the same participant. Sixteen ascending doses of salvinorin A were administered, with four placebo sessions randomly intermixed such that each block of five sessions contained one placebo. Doses were adjusted for bodyweight and are reported in the extracted text as ranging from 0.375 g/kg to 21 g/kg; the extraction does not clearly report whether those units reflect the final intended units. During each session the flask residue was heated underneath with a butane microtorch while the seated participant inhaled slowly for 40 s, followed by a cued exhale. The unblinded session monitor heated the flask behind a divider; a blinded staff member collected ratings and other data. Each participant completed 20 sessions over 8–14 weeks with at least 1 day between consecutive sessions. Before each session participants provided breath and urine samples to screen for recent alcohol and certain drugs; female participants provided pregnancy tests. Sessions lasted about 2.5 h; participants wore eyeshades for 3–5 min before and 10–30 min after dosing and listened to a continuous instrumental music track. Physiological measures (blood pressure and heart rate) were recorded at baseline and every 2 min for 60 min post-inhalation. Subjective drug strength was rated on an 11-point scale (0 = no effect; 10 = strongest imaginable for salvinorin A) at baseline and every 2 min for 60 min; if a participant was unresponsive during early time points the value of 10 was imputed. Approximately 1 h after administration participants completed the Hallucinogen Rating Scale (HRS) and the Mysticism Scale referencing their session experience. Additional retrospective ratings included peak drug strength, drug liking/disliking, and good/bad effects. Tremor was assessed with the Fahn–Tolosa–Marin Tremor Rating Scale at baseline and at 15 and 30 min post-dose. Participants were also asked whether they would refuse to receive the same or higher dose in future sessions as a safety screen. For analysis the time-course drug strength data were summarised as area under the curve (AUC) using the trapezoidal rule for each participant. Repeated-measures ANOVA with within-subjects factor dose (17 levels including placebo) tested dose-related changes in subjective and physiological measures; placebo ratings across four sessions were averaged to form a single placebo data point. Tukey's HSD was used for post-hoc comparisons with a family-wise error rate of α = .05.

Time course: Subjective drug strength showed orderly time- and dose-related effects. Mean ratings peaked at the first post-inhalation assessment (2 min) and declined toward baseline thereafter; by about 20 min after inhalation mean ratings were near a "possible mild" effect. Two participants provided maximal ratings of 10 on one or more occasions; both male participants were unresponsive at certain high-dose time points and those missing early ratings were imputed as 10. Dose–response and questionnaire outcomes: Peak (end-of-session) ratings of drug strength showed a significant effect of dose (F(16,48) = 12.6, p < .001), with post-hoc tests indicating that doses in the range 4.5–21 g/kg (as reported in the extraction) were generally rated stronger than placebo. AUC analyses also produced a significant dose effect (F(16,48) = 4.5, p < .001), though post-hoc comparisons were more conservative and found only the highest two doses significantly greater than placebo, likely reflecting inter-participant variability in duration. Retrospective ratings of drug liking and "good effects" increased with dose (liking: F(16,48) = 7.5, p < .001; good effects: F(16,48) = 4.4, p < .001). Mean ratings at the highest dose indicated generally positive responses: liking M = 3 (range 2–4), good effects M = 2.25 (range 1–4), disliking M = 0.5 (range 0–2), and bad effects M = 0.5 (range 0–1); only one participant reported "disliked somewhat." On the HRS and Mysticism Scale there were significant dose effects on all HRS subscales and on the Mysticism Scale overall (p < .05). HRS–Intensity showed significant increases versus placebo beginning at 4.5 g/kg, whereas other HRS subscales (Cognition, Affect, Somaesthesia, Perception) were elevated primarily at the highest dose levels and Volition did not show a significant dose effect. Mysticism Scale scores were significantly greater than placebo at the three highest dose levels. Safety and tolerability: There were no significant dose-related changes in heart rate or blood pressure (p > .05). No resting or kinetic tremor was observed (TRS = 0 throughout). No participant refused to receive the same or a higher dose at the end of any session, and no adverse events were reported. The small sample and the selection of hallucinogen-experienced healthy volunteers are noted in the extracted text as constraints on generalisability.

Johnson and colleagues interpret the findings as showing that, under the controlled and supportive conditions employed and in a small sample of healthy hallucinogen-experienced adults, inhaled salvinorin A produced rapid-onset, short-duration, and dose-related subjective effects without measurable physiological harm at the tested doses. The investigators note the unusually high subjective intensity observed in some volunteers, with two providing ratings of 10, and propose that the examined dose range will inform dosing choices for future work (they suggest 3, 7.5 and 21 g/kg as low, medium and high doses based on the present results). The authors also indicate that, given the lack of plateau in drug strength and absence of adverse events in this sample, cautious exploration of higher doses may be considered in future studies. In terms of phenomenology, salvinorin A occasioned subjective effects that overlapped with classic hallucinogens on several HRS domains and elicited mystical-type experiences at high doses as measured by the Mysticism Scale. Nevertheless, the investigators report qualitative differences from serotonergic hallucinogens: perceptual and cognitive HRS subscales were of lower magnitude relative to intensity, and the overall profile was described as unique though bearing some similarities to intravenous dimethyltryptamine and oral psilocybin at high doses. The authors acknowledge several limitations. The small sample size and selection of hallucinogen-experienced, healthy participants limit safety conclusions and generalisability. The volatilisation and inhalation method lacked quantitative temperature control and tubing was not replaced between sessions for the same participant, which could theoretically permit deposition and re-volatilisation across sessions; however, visual observations and a post-hoc test comparing first and final placebo session peaks (p = .18) argued against a major impact of re-volatilisation. The investigators also suggest that the relative absence of dysphoric effects—contrasting with prior reports of kappa agonist-induced dysphoria—may reflect participant expectations, experience, and the preparatory supportive setting used in the study. Finally, the authors conclude these preliminary results demonstrate the feasibility of further human research into salvinorin A and related high-efficacy kappa agonists, while emphasising the need for caution and larger, more diverse samples.