Pharmacokinetics and Pharmacodynamics of an Innovative Psychedelic N,N-Dimethyltryptamine/Harmine Formulation in Healthy Participants: A Randomized Controlled Trial

Mescaline is a classic serotonergic psychedelic alkaloid found in several cactus species and, like LSD, psilocybin, and N,N-dimethyltryptamine, acts primarily via 5-HT2A receptor agonism. Prior human data on mescaline are limited: older radiolabel work in a small male sample and few modern controlled studies have left gaps in knowledge about dose–exposure relationships, metabolites, urinary recovery, and how plasma concentrations relate to acute subjective effects. Because mescaline has relatively low 5-HT2A affinity and modest blood–brain barrier permeability, oral doses of several hundred milligrams are typically required, motivating the need for quantitative pharmacokinetic and pharmacokinetic–pharmacodynamic (PK–PD) characterisation across a wide dosing range. Mueller and colleagues set out to characterise the pharmacokinetics, urinary recovery, and PK–PD relationship of oral mescaline hydrochloride (HCl) in healthy adults. Using richly sampled data pooled from two randomized, double-blind Phase I trials with doses from 100 to 800 mg, the investigators applied non-compartmental analyses and compartmental PK and PK–PD modelling to relate plasma mescaline and metabolite concentrations to acute subjective effects assessed with visual analogue scales. The study aimed to provide modern human data to inform dosing, understand first-pass metabolism to the main metabolite 3,4,5-trimethoxyphenylacetic acid (TMPAA), and consider implications for renal elimination.

The analysis pooled data from two previously published randomized, double-blind crossover trials conducted at the University Hospital Basel. Healthy volunteers aged 25–65 years with body mass index 18–29 kg/m2 were eligible; key exclusions included pregnancy, major psychiatric disorders in the participant or a first-degree relative, chronic somatic illness, and interfering medications. After exclusions for early emesis and incomplete urine collection, the final dataset comprised 49 participants (25 female) and 105 complete administration profiles. Study 1 contributed 300 mg and 500 mg mescaline conditions (parallel cohorts of 16 participants each within a larger crossover comparing multiple psychedelics), whereas Study 2 used a within-subject crossover of 100, 200, 400, and 800 mg doses and an 800 mg plus ketanserin (40 mg) condition. Washouts were at least 10 days (Study 1) or 14 days (Study 2). Dosing sessions began at 09:00 after a standardised small breakfast and outcome assessments continued for 24–30 h; participants remained under supervision during acute effects. Mescaline HCl capsules (100 mg units) were produced to Good Manufacturing Practice standards, and the exact analytical content per capsule was used in dosing inputs. Plasma and urine were sampled at frequent, pre-specified timepoints (multiple samples from 0–24 h in Study 1 and 0–30 h in Study 2). Urine was collected in intervals (Study 1) or cumulatively to 30 h (Study 2). Concentrations of mescaline, TMPAA, and N-acetylmescaline (NAM) were quantified by a validated HPLC–tandem mass spectrometry method. Profiles with emesis within 1 h after dosing or within 1 h after observed Cmax were excluded from absorption-related analyses to avoid biased estimates. Acute subjective effects were assessed repeatedly using single-item visual analogue scales (VASs; 0–100%) including "any drug effect" (primary PD outcome), "good drug effect," "bad drug effect," "nausea," and an item assessing "ego dissolution." VASs were collected at the same timepoints as blood draws. Pharmacokinetic analyses comprised non-compartmental analyses (NCA) to derive observed Cmax, tmax, terminal half-life (t1/2), and AUC0–last and AUC∞ (with percentage extrapolation reported). Urinary recovery was computed in molar units and expressed as absolute and percentage recovery; renal clearance was calculated as urinary recovery divided by AUC∞. Sequential compartmental PK and PK–PD modelling was performed in Phoenix WinNonlin 8.4 using actual sampling times. Initial structural PK modelling tested one- and two-compartment models with first-order absorption and elimination; a lag time (tlag) was considered to represent capsule dissolution. Parameter estimation used a naïve pooled engine with model selection informed by visual fits, changes in -2 log-likelihood (-2LL), Akaike information criterion (AIC), parameter precision, and residual diagnostics; likelihood-ratio tests were applied for nested model comparisons (a drop of 3.84 in -2LL for one additional parameter indicated p = 0.05). Below-limit-of-quantification values post-dose were handled using a censored likelihood approach equivalent to the M3 method. For PD modelling, an Emax sigmoid model constrained to an Emax of 100% (consistent with VAS range) was linked to plasma concentrations via an effect compartment with a first-order equilibrium rate constant (ke0) to account for observed hysteresis. The structure fitted first to "any drug effect" and then applied to exploratory PD items; "bad drug effect" and "nausea" were not modelled because of poor correlation with plasma concentrations. Derived PD endpoints included EC50, ke0, γ (steepness), time of onset, Emax, and effect duration using NCA of model-predicted effect–time profiles.

Participant characteristics and data inclusion: Forty-nine healthy participants (mean age 30 ± 7 years; 25 female) contributed to the pooled dataset. Prior psychedelic experience was common (69% of participants, median about five prior exposures). Early vomiting occurred predominantly at higher doses (eight participants: 1 after 400 mg, 5 after 800 mg, and 2 after 800 mg plus ketanserin); after excluding affected profiles, 105 administrations remained for analysis. Non-compartmental and urinary recovery findings: Over the sampled intervals (24–30 h), a geometric mean of 53% of the administered mescaline dose was recovered unchanged in urine, while TMPAA accounted for about 31% of the dose. Recovery was nominally lower in Study 1 (24 h sampling) than Study 2 (30 h sampling), and cumulative recoveries in some Study 2 cases approached near-complete dose recovery. More than 50% of recovered mescaline, TMPAA, and NAM appeared in urine within the first 8 h post-dose. Both mescaline and TMPAA demonstrated dose linearity between 100 and 800 mg. NAM concentrations were low (peak and AUC <10% of mescaline), with a shorter t1/2 (~2 h), suggesting limited clinical relevance. Compartmental PK model: A one-compartment model with first-order absorption and elimination plus a lag time provided the best balance of fit and parsimony. Inclusion of a second compartment yielded statistically significant -2LL drops in some individual profiles but introduced strong parameter correlations and did not improve visual fit, so it was discarded. Adding tlag significantly improved fit in 59/105 profiles (mean improvements of -2LL = 11 ± 12 and AIC = 8.7 ± 12 across profiles). Key PK descriptors included rapid absorption with a geometric mean tmax of ~2.0 h and a mean lag time of ~0.2 h; terminal plasma half-lives were ~3.5 h across doses. Apparent clearance (Cl/F) averaged about 42 L/h; assuming ~50% oral bioavailability (inferred from urinary recovery), this corresponds to an estimated total clearance near 21 L/h. PK–PD modelling and subjective effects: The sigmoid Emax model with an effect compartment (ke0) was necessary to capture the counterclockwise hysteresis between plasma concentrations and subjective effects. Adding a steepness parameter (γ) improved fit in 81/91 PD profiles evaluated, and inclusion of ke0 further improved fit in 74/91 profiles (mean -2LL improvement 30 ± 25). Model-predicted "any drug effect" typically began around 1 h after dosing and peak effects occurred between about 1.9 h (100 mg) and 4.2 h (500 mg) in the profiles reported. Maximal effect magnitude and duration were dose-dependent: at 100 mg the model predicted an Emax of ~13% with an effect duration ~2.8 h, whereas at 800 mg predicted Emax was ~89% with a duration of ~15 h. "Good drug effect" and "ego dissolution" were modelled similarly in exploratory analyses; "bad drug effect" and "nausea" did not correlate well with plasma concentrations and were not modelled. Coadministration of ketanserin with 800 mg mescaline markedly attenuated subjective response and reduced mescaline-induced emesis, and emesis at the highest dose may have reduced effective absorption in some sessions. Overall, PK model predictions aligned well with NCA-derived metrics.

Mueller and colleagues interpret their results as the first modern, detailed characterisation of oral mescaline HCl pharmacokinetics, PK–PD relationships, and urinary recovery across 100–800 mg in healthy volunteers. The data show dose-proportional increases in AUC and Cmax over this range and indicate that a simple one-compartment model with first-order absorption and elimination plus a small lag time adequately describes plasma concentrations. Rapid absorption (median tmax ~2.0 h) and a plasma half-life of ~3.5 h were consistently observed. The investigators attribute similar early tmax and Cmax values for mescaline and its main metabolite TMPAA to substantial first-pass metabolism, estimating that roughly half the oral dose is converted to TMPAA before systemic exposure, and they estimate oral bioavailability of at least 53% based on urinary recovery. The authors highlight that mescaline and TMPAA exhibit parallel concentration–time profiles and that renal excretion appears to be the dominant elimination pathway for both analytes; measured renal clearance averaged ~22 L/h (366 mL/min), higher than typical glomerular filtration rates and potentially indicative of active tubular secretion. In terms of pharmacodynamics, subjective effects were tightly linked to plasma concentrations but displayed a concentration–effect delay (counterclockwise hysteresis) that was accommodated by an effect compartment (ke0), consistent with delayed central distribution. The PK–PD relationship resembled prior findings for LSD and psilocybin and supports the concept that 5-HT2A receptor agonism, when accompanied by sufficient plasma/brain exposure, underlies acute subjective effects. Strengths noted by the investigators include the large number of richly sampled profiles across a broad dose range, analytically confirmed dosing, controlled clinical conditions, and combined PK and urinary recovery data enabling an indirect bioavailability estimate. Several limitations were acknowledged: the healthy, mostly young participant sample limits generalisability to older adults or patients with impaired renal function; prior psychedelic experience in many participants could have influenced PD responses; methodological differences between the two studies (notably sampling duration) may have affected total recovery estimates; and limited early sampling could blunt precision in absorption-phase parameter estimates and the lag-time assessment. The authors also note modelling limitations: extremely low-effect PD profiles produced unrealistically high EC50 estimates that were handled by exclusion from some summaries, and a population-based PK–PD approach might better characterise intra- and interindividual variability. Finally, they affirm that an intravenous mescaline study would be required to confirm the bioavailability estimate.

Mescaline HCl displayed dose-proportional pharmacokinetics from 100 to 800 mg and was well described by a one-compartment model with first-order absorption and elimination. Acute subjective effects tracked plasma concentrations with a short response lag and were described by a sigmoid Emax model without evidence of acute tolerance. The data are consistent with an oral bioavailability of at least 53% due to first-pass conversion to TMPAA, and renal excretion appears to be the primary elimination route for both mescaline and TMPAA.