High‑dose inhaled salvinorin A produced peak plasma concentrations within 2 minutes followed by a rapid decline, and these plasma levels were significantly correlated with participant- and monitor-rated drug effects—this is the first human study to demonstrate a direct pharmacokinetic–effect relationship. Prolactin rose from 5 minutes and peaked at 15 minutes while cortisol responses were inconsistent, and the results confirm the efficacy of the vapour inhalation technique.
Salvinorin A is a kappa opioid agonist and the principal psychoactive constituent of the Salvia divinorum plant, which has been used for hallucinogenic effects. Previous research on salvinorin A pharmacokinetics likely underestimated plasma levels typically resulting from the doses administered due to inefficient vaporization and not collecting samples during peak drug effects. Six healthy adults inhaled a single high dose of vaporized salvinorin A ( n = 4, 21 mcg/kg; n = 2, 18 mcg/kg). Participant- and monitor-rated effects were assessed every 2 min for 60 min post-inhalation. Blood samples were collected at 13 time points up to 90 min post-inhalation. Drug levels peaked at 2 min and then rapidly decreased. Drug levels were significantly, positively correlated with participant and monitor drug effect ratings. Significant elevations in prolactin were observed beginning 5 min post-inhalation and peaking at 15 min post-inhalation. Cortisol showed inconsistent increases across participants. Hormonal responses were not well correlated with drug levels. This is the first study to demonstrate a direct relationship between changes in plasma levels of salvinorin A and drug effects in humans. The results confirm the efficacy of an inhalation technique for salvinorin A.
Salvia divinorum has been used historically in Mazatec shamanic practice and, more recently, has become a recreational psychoactive substance. Its principal psychoactive constituent, salvinorin A, is a kappa opioid receptor agonist that produces hallucinogenic effects distinct from classic serotonergic hallucinogens because it is not active at 5-HT2A receptors. Previous human pharmacokinetic data are limited and potentially incomplete: an earlier inhalation study measured plasma levels at only three post-inhalation time points that occurred well after the very rapid subjective peak observed in laboratory work, and differences in vaporisation techniques may have produced underestimates of typical plasma exposure from smoked or vapourised doses. Johnson and colleagues set out to characterise more precisely the time course of salvinorin A plasma levels after inhalation using a relatively efficient vapour-delivery method, and to relate those plasma levels to concurrent subjective and monitor-rated drug effects. The investigators also measured prolactin and cortisol, hormones known to respond to kappa agonists, and assayed residual drug remaining in the glass pipe to assess delivery efficiency. Frequent early blood sampling was used so that peak drug levels and their temporal relation to subjective effects could be examined directly.
Papers cited by this study that are also in Blossom
Johnson, M. W., Maclean, K. A., Reissig, C. J. et al. · Drug and Alcohol Dependence (2011)
Maclean, K. A., Johnson, M. W., Reissig, C. J. et al. · Psychopharmacology (2012)
Maqueda, A. E., Valle, M., Addy, P. H. et al. · International Journal of Neuropsychopharmacology (2015)
Papers in Blossom that reference this study
This report analysed data from six healthy adults who had participated in a prior laboratory study of inhaled salvinorin A; the same participants completed a final, single high-dose inhalation session that included blood sampling. The mean age was 25 years (range 21–35). Participants reported prior S. divinorum use (mean 11 occasions) and prior classic hallucinogen use (mean 32 occasions). Two other individuals from the earlier cohort did not participate in the blood-draw session for safety or personal reasons. Each participant inhaled a single high dose of vapourised salvinorin A chosen as their highest tolerated dose from prior ascending-dose sessions: four participants received 21.0 mcg/kg and two received 18.0 mcg/kg. Subjective drug strength and a set of monitor-rated effects (including distance from usual daily reality, unresponsiveness, psychological distress, paranoia, anxiety/fear, motor activity, joy/peace, and physical distress), together with physiology (systolic and diastolic blood pressure and heart rate), were assessed every 2 minutes for 60 minutes after inhalation. Blood was collected at 13 time points up to 90 minutes post-inhalation; plasma was obtained by cold centrifugation. Salvinorin A was quantified by solid-phase extraction followed by LC‑MS/MS in triplicate, using a +5 mass analogue as internal standard. Prolactin and cortisol were assayed by ELISA in triplicate and averaged. Residual salvinorin A in the glass pipe was measured by three dichloromethane washes, drying, reconstitution and LC‑MS/MS analysis (triplicate samples). For statistical analysis, Pearson correlations were computed within each participant between hormone assays and subjective/monitor ratings using the seven time points common to blood draws and ratings (baseline, 2, 4, 10, 20, 30, and 60 minutes). Correlations between salvinorin A and hormones used the same seven time points, while correlations among hormones used all 13 time points. Group-level peak-value correlations (using each participant’s maximal value across the time course) were also examined to allow for delayed hormonal responses. Repeated-measures regression (SAS PROC MIXED with an AR(1) covariance structure) modelled the relationship between salvinorin A plasma level and participant and monitor ratings from baseline to 60 minutes; significance was set at p < .05. The percent of the prepared dose remaining as residue in the glass pipe was calculated from residual mass and the prepared absolute dose. The extracted text notes insufficient sample volumes for two hormone time points for one participant (missing 1 and 10 minute prolactin/cortisol results).
Plasma assays for salvinorin A were successfully obtained at all planned time points; triplicate assay coefficients of variation (CVs) for plasma samples and standards were < 10%. Mean salvinorin A levels peaked at 2 minutes post-inhalation and then declined rapidly, with more gradual reductions thereafter until levels approached baseline by 90 minutes. Individual peak times varied, occurring as early as 1 minute and as late as 4 minutes in some participants. Statistical analyses showed that salvinorin A plasma levels were significantly, positively associated with participant-rated drug strength and with certain monitor-rated drug effects across the time course. In contrast, salvinorin A level did not significantly predict a set of other monitor ratings (anxiety/fear, motor activity, joy/peace, physical distress) nor physiological measures (systolic and diastolic blood pressure, heart rate). These results were unchanged after controlling for lifetime use of classic hallucinogens and lifetime use of S. divinorum. Prolactin showed a reliable increase following salvinorin A administration, with mean peak levels occurring at about 15 minutes post-inhalation; some individuals exhibited a plateau of elevated prolactin from roughly 10 to 30 minutes. Cortisol showed a superficially similar mean time course but large individual variability, and some participants exhibited little evidence of a cortisol response. Within individual participants there were no significant correlations between cortisol or prolactin levels and drug effect ratings; the only notable individual correlations with physiology were one positive correlation between pulse and cortisol, and one negative correlation between systolic blood pressure and prolactin. Salvinorin A levels were not significantly correlated with either cortisol or prolactin within any individual. Prolactin and cortisol were positively correlated within participants across the full set of time points (Pearson r range .36–.92), with this correlation reaching significance in four of six participants. At the group level, correlations between peak values detected no significant relation between salvinorin A and prolactin (p = .82), salvinorin A and cortisol (p = .15), or prolactin and cortisol (p = .68). Residual salvinorin A recovered from the glass pipe averaged 57.1 mcg (SD = 24.3 mcg), representing a mean of 4.21% (SD = 2.25%) of the prepared absolute dose; CVs for triplicate residual assays were < 4%.
Johnson and colleagues emphasise three main findings: the vapour-delivery method was comparatively efficient; salvinorin A blood levels rose and fell rapidly with mean peak at about 2 minutes; and hormonal responses—prolactin reliably and cortisol variably—followed a delayed and more prolonged time course compared with the drug’s plasma profile. The investigators note that mean peak plasma levels in this study were substantially higher than reported in a prior human inhalation study despite lower absolute administered masses, which they attribute to greater delivery efficiency of the glass-pipe vapourisation method; factors such as temperature and airflow topography are suggested as possible contributors to that efficiency. In terms of pharmacodynamics, the study demonstrated a strong temporal correspondence between concurrent plasma salvinorin A levels and participant- and monitor-rated subjective drug effects, supporting the view that subjective intensity is closely tied to contemporaneous systemic exposure. Hormonal measures behaved differently: prolactin increased reliably but with a lag relative to plasma drug levels, and cortisol responses were inconsistent across participants. The authors interpret these hormonal patterns as evidence that endocrine responses to kappa agonism may be temporally dissociated from immediate subjective effects. The study team also highlights the practical relevance of the findings: because smoking and vapourisation both use inhalational routes, the pharmacokinetic data and the evidence for efficient delivery have implications for understanding recreational S. divinorum use and, to a limited extent, for potential therapeutic applications delivered via inhalation. The discussion presents individual variability in peak timing and hormonal responses as notable observations. The extracted text does not provide an extended list of limitations beyond reporting some missing hormone samples and inter-individual variability, so no additional author-acknowledged caveats are introduced here.
The investigators conclude that an efficient vapourisation method produced substantially higher salvinorin A plasma levels than previously reported, and that subjective drug intensity closely tracked concurrent plasma concentrations. Prolactin generally increased after administration but on a delayed and more prolonged time course than the drug itself, while cortisol increases were inconsistent across participants. These findings improve understanding of the pharmacokinetics and short-term hormonal effects of inhaled salvinorin A and are relevant to recreational use of S. divinorum and to considerations of inhalational delivery in any future therapeutic contexts.
The plant Salvia divinorum (a member of the mint family) has been used historically in shamanic practices of the Mazatec people of Oaxaca, Mexico for at least several hundred years, although it was not botanically described until the 1960s. Within the past 15 years S. divinorum has gained increased popularity as a psychoactive drug in non-traditional contexts. In non-traditional use, products containing S. divinorum leaves, sometimes infused with S. divinorum extract in order to increase drug effects, are typically smoked. Salvinorin A, the primary psychoactive compound in S. divinorum, is a kappa opioid agonist hallucinogen that is not active at the 5-HT2A receptor, the primary site of activity for classic hallucinogens such as lysergic acid diethylamide (LSD) and psilocybin. Although S. divinorum and salvinorin A have not been controlled at the federal level in the US, at the time of this writing at least 35 states within the US and 27 nations have enacted various levels of restriction for S. divinorum. Understanding the effects of salvinorin A, including its pharmacokinetic profile, in humans is important for understanding recreational use of S. divinorum. Laboratory research has not found evidence of persisting psychotic-type episodes resulting from salvinorin A. Cases of persisting psychotic-type episodes have been reported in association with recreational use, although the causal role of S. divinorum remains unclear. In addition, the dissociative and perceptual effects resulting from salvinorin Acould potentially result in dangerous behavior in an unsupervised environment. Therefore, studying the pharmacokinetic profile of salvinorin A may inform the understanding of potential adverse reactions observed in recreational S. divinorum use. Examining human salvinorin A effects is also important because salvinorin A or derivative compounds may serve as therapeutic agents for neurological (e.g. Alzheimer's disease), pain, mood, personality, gastrointestinal, and cocaine-use disorders. While a study of inhaled salvinorin A would primarily model inhaled recreational use of S. divinorum, its results may also have limited relevance for potential therapeutic applications. Regardless of whether potential therapeutic applications would deliver the drug via vaporization, examining the relation between plasma drug levels and resulting subjective effects in the vaporized route may inform the more general relation between plasma drug levels and subjective effects at play in potential therapeutic applications. Studies of intraperitoneally injected salvinorin A in rats, and intravenously injected salvinorin A in rhesus monkeys, have provided basic pharmacokinetic data. However, cross-species differences and differences between routes of administration may limit the implications of these findings for the pharmacokinetics of inhaled salvinorin A in humans. One previous study assessed the pharmacokinetic profile of salvinorin A in humans. That study showed increases in salvinorin A, prolactin, and cortisol resulting from inhaled administration of the drug. However, there are several issues that remain unexamined. First, the previous study did not examine the relation between individual salvinorin A pharmacokinetic and psychoactive effects. Second,examined plasma levels of salvinorin A at only 3 time points post-inhalation (15, 20, and 30 min), with the first assessment occurring substantially after the time at which peak participant-rated effects were observed in our research (2 min post-inhalation), suggesting that pharmacokinetic analysis of peak drug effects was missed. Third,used a commercial vaporizer to deliver a maximum dose (12 mg) that was approximately eight to twelve times higher than the maximum doses administered in studies that used a glass pipe to vaporize salvinorin Aand a study that had participants smoke S. divinorum leaves infused with additional salvinorin A. Substantial differences with the previously reported pharmacokinetic study regarding dose and delivery system warrant a pharmacokinetic analysis of our study. In the present study, we examined the time course of salvinorin A plasma levels after inhalation of a high dose, delivered via a relatively efficient vaporization system. Blood was drawn at relatively frequent post-inhalation time points in order to accurately describe plasma levels surrounding the relatively rapid peak drug effects of salvinorin A. This frequent sampling allowed us to analyze the correspondence between drug levels and subjective effects throughout the drug time course. In addition, we examined levels of prolactin and cortisol, which are both sensitive to kappa agonist administration. In order to address one aspect of the efficiency of the delivery system, residual salvinorin A from the glass pipe was assayed for each session.
Participants were 6 individuals who participated in a previous study assessing the effects of inhaled salvinorin A in the laboratory. The sample size was judged sufficient for examining pharmacokinetic data because robust significant subject-rated effects were observed with fewer participants. Participants had taken part in up to 20 previous sessions (16 salvinorin A doses in ascending order and 4 intermixed placebo sessions under blind conditions) that did not involve collecting blood samples. Two individuals (one female and one male) whose subjective and cognitive data were included in our previous sample of eightdid not participate in the final salvinorin A administration session, which was the only session involving blood draws. In the case of the male, the participant decided not to participate in the blood draw session upon considering several subjectively intense sessions previously in the study. In the case of the female, the investigators decided not to continue her onto the blood draw session due to excessive spontaneous arm movements in previous sessions, which may have interfered with the blood draws. For the six participants reported here, the mean age was 25 years (range: 21-35). They reported using S. divinorum on a mean of 11 previous occasions (range: 1-40), with their reported first use at a mean age of 21 years (range: 16-31). They reported using classic hallucinogens on a mean of 32 previous occasions (range: 5-111). Study staff who were present during drug administration had established a rapport with participants during previous preparatory sessions and lower dose and placebo sessions as described previously.
Each participant inhaled a single high dose of vaporized salvinorin A. The dose administered was the highest tolerated dose of salvinorin A in previous sessions. For four participants, this dose was 21.0 mcg/kg, which was the maximal dose in the dose run-up. For the other two, this dose was 18.0 mcg/kg because they replied "yes" to a question asking them if they would refuse to receive the same or higher doses at the conclusion of a 19.5 mcg/kg session. As described previouslysubjective drug strength and monitor-rated effects (drug strength, distance from usual daily reality, unresponsiveness, psychological distress, paranoia, anxiety/fear, motor activity, joy/peace, and physical distress) and physiology measures (systolic and diastolic blood pressure and heart rate) were assessed every 2 min for 60 min after inhalation. Blood samples were collected at 13 time pointsand cold centrifuged to obtain plasma. Plasma samples were purified by solid phase extraction and analyzed in triplicate via liquid chromatography-tandem mass spectrometry (LC-MS/MS) using a +5 mass analogue of salvinorin A as the internal standard. Analyses of prolactin and cortisol were performed with enzyme-linked immunosorbent assay (ELISA) kits (Calbiotech, Spring Value, CA) and were run in triplicate. The average of these three assays was used in the analyses. Residual salvinorin A in the glass pipe was determined for each session. Specifically, dichloromethane (1 mL in three separate washes) was used to wash the inner surfaces of the glass pipe. The combined 3 mL of resulting solution was then dried under a stream of nitrogen. Three separate samples from the resulting residue were dissolved into mobile phase and subjected to LC-MS/MS for analysis.
For each participant, we calculated Pearson's correlations between each hormonal assay (prolactin and cortisol) and subjective and monitor ratings of drug effects, using only the 7 time points when both blood and ratings were collected (baseline, 2, 4, 10, 20, 30, and 60 min post-inhalation). Correlations were also conducted for each participant between salvinorin A and prolactin levels, between salvinorin A and cortisol levels, and between prolactin and cortisol levels. Correlations between drug and hormones used the 7 common time points indicated above, while correlations among hormones used 13 common time points. Because a delayed hormonal response (relative to drug levels) might obscure a relationship between drug and hormonal levels, the same pairs of correlations were also conducted at the group level using peak values (i.e. single maximal value across the time course for each participant) for drug and hormonal levels. Repeated measures regression (SAS PROC MIXED, AR(1) covariance structure) was used to model the relationship between salvinorin A plasma level and participant and monitor ratings of drug effects from baseline to 60 min post-administration. As with the correlations, this analysis only used the 7 time points common to both blood draws and drug strength ratings. Statistical significance was defined as p < .05. The percent of the intended dose that remained as residual salvinorin A in the glass pipe was calculated using the salvinorin A residual mass for each participant (i.e. mean of the triplicate LC-MS/MS assays) and the prepared absolute salvinorin A dose for each participant (i.e. taking bodyweight into account).
Samples were collected and assayed for salvinorin A level at all time points. For prolactin and cortisol, sample volumes were insufficient to obtain results for two time points for one participant (at the 1 and 10 min time points). Coefficients of variation (CV) for plasma samples (in triplicate) and standards (in duplicate) were < 10%. Figure(a) shows mean salvinorin A levels at all blood collection time points (up to 90 min post-inhalation). In order to show individual variability contributing to mean levels, Figure(b) shows individual participant salvinorin A levels at each time point up to 30 min post-inhalation. Figureshows that mean peak salvinorin A levels occurred at 2 min post-inhalation, followed by rapid reductions and then more gradual reductions until the final time point at 90 min post-infusion, at which time salvinorin A levels were close to baseline (zero). Although these trends were generally observed at the individual participant level (Figure), notable variations occurred, with peak effects occurring as early as 1 min to as late as 4 min post-inhalation. To illustrate the relationship between drug blood levels and subjective drug strength, Figure). The effect of salvinorin A level was not significant for the remaining monitor ratings (anxiety/fear, motor activity, joy/peace, and physical distress). The effect of salvinorin A level was also not significant for physiology measures (systolic and diastolic blood pressure and heart rate). Results remained unchanged after controlling for lifetime use of hallucinogens and lifetime use of S. divinorum. Figureshows the effects of salvinorin A administration on plasma prolactin. Figure(a) shows mean prolactin levels, and Figure(b) shows prolactin levels in individual participants. Mean peak effects occurred at 15 min post-inhalation and gradually decreased through 90 min. However, individual participant data show a plateau of peak prolactin levels from 10 to 30 min post-inhalation for some individuals. Figureshows the effects of salvinorin A administration on plasma cortisol. Figureshows mean cortisol levels, and Figure(b) shows cortisol levels in individual participants. The mean cortisol time course resembled that of prolactin. However, there was substantial individual variability with little evidence of a cortisol response observed in some participants. There were no significant correlations between cortisol or prolactin levels and drug effect ratings within individual participants. No individual participant correlations between hormone levels and physiological measures were significant with the exception of one positive correlation between pulse and cortisol and one negative correlation between systolic blood pressure and prolactin. Salvinorin A levels were not significantly correlated with either cortisol or prolactin levels within any individual participant. Levels of prolactin and cortisol were positively correlated within each participant across the 13 time points (1 participant with 11 time points due to missing data) (Pearson r range: .36-.92; significant for 4 of 6 participants). In correlations at the group level, no significant relation was detected between salvinorin A and prolactin levels (p = .82), between salvinorin A and cortisol levels (p = .15), or between prolactin and cortisol levels (p = .68). CV for the triplicates of residual salvinorin A assays for each participant were < 4%. The mean mass of salvinorin A residue in the glass pipe across participants was 57.1 mcg (standard deviation (SD) = 24.3 mcg), representing a mean of 4.21% (SD = 2.25%) of the prepared absolute dose.
This study is unique in that it examined the time course (including frequent, early time points) of salvinorin A plasma levels after salvinorin A inhalation, delivered via a relatively efficient vaporization system. The present study resulted in novel information relevant to three domains: drug delivery, time course of drug levels, and time course and magnitude of hormonal effects.
The present study showed substantially higher salvinorin A plasma levels compared to the previous study of inhaled salvinorin A pharmacokinetics. The previous study found a mean salvinorin A level of approximately 0.9 to 1.0 ng/mL resulting from 8 and 12 mg salvinorin A (with little difference between those two doses). In contrast, in the present study, at doses approximately eight times lower (18.0 and 21.0 mcg/kg, which equate to ~1.26 and 1.47 mg for a 70 kg bodyweight person), resulted in a mean of 18.8 ng/mL at peak effects. These data suggest the present study used a substantially more efficient delivery method. Differences in efficiency could involve multiple factors including temperature and air flow topography. Moreover, the analysis showing only a small percentage of residual salvinorin A in the glass pipe highlights the efficiency of the delivery system.
The present study found strong correspondence between salvinorin A levels and ratings of drug strength throughout the time course. Unlike the previous study of salvinorin A pharmacokinetics, this study was able to demonstrate this relationship due to more frequent drug effect rating assessments and blood draws. The present results indicate that subjective effects of salvinorin A are a direct function of concurrent plasma levels of the drug. This finding is consistent with a study of intravenous salvinorin A in rhesus monkeys reporting overt sedation-like behavior effects generally overlapping with the period of detected plasma levels of salvinorin A (e.g. within ~15 min post-injection).
Similar to, the present study showed increases in prolactin and, less consistently, cortisol following salvinorin A administration. Due to infrequent sampling, the previous study did not have the ability to determine how closely hormone levels and salvinorin A levels were related in time. By showing rapid increases in salvinorin A levels that match the rapid subjective effects of the drug, the present study had the potential to demonstrate a strong correspondence between drug and hormone levels. However, the present study showed that prolactin and cortisol responses to salvinorin A administration followed a more delayed and prolonged time course than the drug itself.
This study provides important information regarding the pharmacokinetics of a relatively novel drug used for its hallucinogenic effects. It confirmed that a relatively efficient vaporization method resulted in substantially higher drug plasma levels compared to a previous study of salvinorin A pharmacokinetics. Moreover, this study showed strong correlations between salvinorin A blood levels and drug strength ratings across the time course of drug effects, suggesting that subjective effects are a product of concurrent blood levels. This study also showed that salvinorin A generally increased prolactin, although it followed a more delayed and prolonged time course than the drug itself. Cortisol showed inconsistent increases across participants. Because smoking and vaporization both involve inhalation, the results of this study may be relevant to the recent use of S. divinorum in non-traditional contexts.
Create a free account to open full-text PDFs.
Garcia-Romeu, A., Kersgaard, B., Addy, P. H. · Experimental and Clinical Psychopharmacology (2016)