Pharmacokinetics of Escalating Doses of Oral Psilocybin in Healthy Adults

Psilocybin is a naturally occurring psychedelic tryptamine that is rapidly dephosphorylated in vivo to the active compound psilocin, which mediates psychoactive effects primarily via serotonin 5‑HT2A receptor agonism. Renewed clinical interest in psilocybin arises from trials showing potential therapeutic effects in anxiety and depression, including single-dose and multi-dose studies in terminal illness and treatment‑resistant depression. Previous human pharmacokinetic (PK) studies of oral psilocybin were small, sampled for relatively short durations, and used assays with higher limits of quantitation, leaving questions about the full 24‑hour PK profile, renal excretion, metabolite handling, and the influence of covariates such as renal function and body weight. Brown and colleagues designed this open‑label, sequentially escalating dose pharmacokinetic study in healthy adults to characterise psilocybin/psilocin PK under contemporary cGCP/cGLP conditions. The primary aim was to build a population PK model for psilocin after oral psilocybin administered at 0.3, 0.45, and 0.6 mg/kg. Secondary aims included assessment of covariates potentially predictive of PK behaviour (for example renal function and weight), determination of renal excretion of psilocin, safety characterisation, and simulation of fixed oral doses to evaluate whether a fixed dosing strategy (e.g. 25 mg) could approximate weight‑based exposures used in prior studies.

This single‑site, open‑label pharmacokinetic study recruited 12 healthy adults over a 15‑month period after institutional review board approval. Eligibility required at least one prior substantial psychedelic experience, absence of recent psychedelic use (no use within 1 month), negative urine drug screens on dosing days, and no current or recent (past 5 years) major psychiatric diagnoses or first‑degree relatives with bipolar or psychotic disorders. Tobacco users (unless on nicotine patch), those requiring medications during the 8‑hour drug action window, and individuals on antidepressants or monoamine oxidase inhibitors were excluded. Each participant received 6–8 hours of preparatory counselling and attended supervised 8‑hour dosing sessions with a consistent male–female guide dyad; debriefing and integration meetings were held the morning after each dose. Psilocybin (synthetic) was formulated into opaque capsules adjusted for water content to deliver individual doses of 0.3, 0.45, or 0.6 mg/kg using actual body weight. Doses were administered at minimum 4‑week intervals. On dosing mornings subjects had baseline ECG, vital signs, pregnancy and urine drug testing, intravenous catheter placement, and a standardised breakfast before ingesting the capsule with 360 mL water. Blood for plasma psilocybin and psilocin was collected predose and at 0.25, 0.5, 0.75, 1, 2, 3, 4, 6, 8, 12, 18 and 24 h postdose. Urine was collected over 24 h using a commode insert and stabilised with ascorbic acid to a final 25 mM concentration before freezing. Plasma and urine were stored at −70 °C until assay. Analysis of psilocybin and psilocin used a validated LC‑MS/MS assay with solid‑phase extraction and HPLC separation; no parent psilocybin was detected in plasma or urine. The lower limits of quantitation (LLOQ) were 0.5 ng/mL for plasma psilocin and 5.0 ng/mL for urine psilocin. Pharmacokinetic evaluation employed both noncompartmental methods (WinNonlin) to obtain Cmax, Tmax and AUC0–24, and population PK modelling (NONMEM) to characterise structural models, intersubject variability and covariate effects. Below‑limit data were handled using the M3 method where applicable. Model development tested one, two and three compartment structures, evaluated covariates including weight and measured creatinine clearance, and considered a reversible psilocin‑glucuronide pool to account for apparent biexponential decay. Model evaluation included 1,000 bootstrap runs and visual predictive checks (VPC) stratified by dose. Renal clearance of psilocin was calculated as the amount excreted in 24‑h urine divided by the plasma AUC0–24. Finally, the final model was used to simulate fixed single doses (20 and 25 mg) with 500 iterations to compare expected exposures against those observed after a 0.3 mg/kg dose. Adverse events were recorded from enrolment through 30 days post‑last dose and graded per CTCAE v4.

The evaluable dataset comprised 353 measurable plasma psilocin concentrations from the 12 participants; inclusion of values below LLOQ using the M3 approach added 48 additional plasma samples. One participant was replaced because no postdose blood samples could be obtained, one received only a single dose due to transient ‘white‑coat’ hypertension, and another received two doses only because of scheduling difficulties. Population modelling identified a one‑compartment model with first‑order absorption and linear clearance as the core structural model, but model fit improved by including a bidirectional compartment representing reversible formation and hydrolysis of psilocin glucuronide. For subjects where a two‑compartment tissue model seemed to fit, the peripheral volume estimate was physiologically implausible, supporting the glucuronide exchange interpretation. The final model applied between‑subject variability (ETA) to clearance (CL/F) and central volume (V/F); attempts to estimate Ka and the glucuronide kinetics directly were limited by the sample size and lack of measured psilocin‑glucuronide concentrations. Inclusion of covariates such as weight, bilirubin, albumin or measured creatinine clearance did not lead to robust, retained predictors in the bootstrap evaluation, and the final model contained no covariates. Noncompartmental analyses supported linear pharmacokinetics of psilocin across the twofold oral psilocybin dose range (0.3–0.6 mg/kg): dose‑normalised AUC and Cmax did not show significant dose dependence. Renal excretion of intact psilocin was minimal: 1.7% of the administered psilocybin dose was recovered as psilocin in urine (95% CI 1.4–1.9%), corresponding to a renal clearance of roughly 1 mL/min/kg, or 58% of measured creatinine clearance. Measured creatinine clearance did not improve the NONMEM model fit. Safety data indicated no serious adverse events; commonly observed effects were mild, transient hypertension and tachycardia, and headaches in the second 12 h that responded to acetaminophen. No cases of hallucinogen persisting perception disorder were reported during study or follow‑up. Simulations (N = 500) using the final model showed that a fixed 25 mg oral dose of psilocybin produced psilocin AUC and Cmax distributions that largely overlapped with those observed after a 0.3 mg/kg dose in this cohort. Visual predictive checks and boxplots indicated that the interquartile ranges for AUC and Cmax following 25 mg approximated those after 0.3 mg/kg, although outliers were influenced by a single 120 kg subject.

Brown and colleagues interpret their findings as confirming rapid dephosphorylation of psilocybin to psilocin, since no parent compound was detectable in plasma or urine. The PK of psilocin over the 0.3–0.6 mg/kg oral psilocybin range behaved linearly in this sample, and the modelling results indicate that reversible glucuronidation of psilocin could plausibly account for an apparent biexponential decay in some subjects. The study team emphasised that they did not measure psilocin‑O‑glucuronide or other reported metabolites, a limitation that prevents direct confirmation of the glucuronide hypothesis and leaves the activity and excretion pathways of those metabolites unresolved. The low fraction of psilocin renally excreted (less than 2%) led the investigators to conclude that no psilocybin dose adjustment appears necessary for patients with mild to moderate renal impairment, and that measured creatinine clearance was not a useful covariate for predicting psilocin clearance in their model. They cautioned, however, that this conclusion presumes that psilocin metabolites are neither active nor substantially renally excreted, which remains untested here. Key limitations acknowledged by the authors include the small sample size, limited representation of women (two female participants) and non‑White subjects (two non‑White participants), and the narrow twofold dose range, which constrained ability to detect nonlinearity or demographic effects. Additional constraints were the lack of assays for glucuronide or other metabolites and limited power to resolve absorption and glucuronide kinetics. In terms of practical implications, the authors note that standardising a cGMP psilocybin product will favour fixed dosing over weight‑based dosing; their simulations suggested that a fixed 25 mg oral dose approximates exposures from a 0.3 mg/kg dose in this population. Finally, they indicated that further reports will examine relationships between plasma psilocin concentrations and psychological effects, and that larger, more diverse studies and cGMP formulations will be required to support later‑phase clinical development.

The study found that psilocin exposure after oral psilocybin was linear across the 0.3–0.6 mg/kg dose range in healthy adults and that reversible formation and hydrolysis of a psilocin glucuronide could explain biexponential decay observed in some participants. All doses were well tolerated in this small cohort, with no serious adverse events. Renal excretion of intact psilocin was minimal, suggesting no dose adjustment is necessary for mild to moderate renal impairment. Finally, simulations indicated that a fixed 25 mg oral dose is expected to produce psilocin AUC and Cmax similar to a 0.3 mg/kg weight‑based dose in this study sample.