Pharmacokinetics and pharmacodynamics of lysergic acid diethylamide microdoses in healthy participants

The study employed a double-blind, randomized, placebo-controlled crossover design with four 6-hour test sessions per participant, each separated by at least 5 days. Sessions began at 09:00 and a single oral dose (placebo, 5, 10 or 20 µg LSD base) was administered at 10:00. The trial was registered in the Dutch Clinical Trial Register (NTR7102). Healthy volunteers aged 18–40 years with body mass index 18–28 kg/m2 and at least one prior hallucinogen experience were recruited from a university setting. Exclusion criteria included pregnancy or lactation, current or past psychiatric disorders, family history of major psychiatric disorders, drug addiction, previous serious adverse reactions to psychedelics, and significant physical illness. The final analysed sample comprised 23 completers (12 male, 11 female; mean age 23 ± 3 years). LSD base was prepared as an oral ethanol solution (25 µg/ml) and aliquots were diluted to give final 1 ml doses containing 5, 10 or 20 µg; placebo was 1 ml ethanol. The authors note conversion factors between LSD base and LSD tartrate used in prior studies (approximately 1 µg base ≈ 1.23–1.33 µg tartrate). Pharmacokinetic sampling used an indwelling catheter with plasma collected before dosing and at 0.5, 1, 1.5, 2, 3, 4 and 6 h. Plasma LSD was measured by ultra-high-performance liquid chromatography–tandem mass spectrometry with a lower limit of quantification of 2.5 pg/ml after a re-extraction procedure. Non-compartmental analyses provided observed Cmax and Tmax and terminal elimination rate estimates; compartmental modelling used a one-compartment model with first-order input and elimination in Phoenix WinNonlin. For pharmacokinetic–pharmacodynamic (PK–PD) modelling, predicted concentrations from the PK model were linked to effect-site concentrations using a first-order equilibrium rate constant (keo) and a sigmoid Emax model (parameters EC50, Emax, γ), with VAS scores bounded 0–10. Subjective effects were assessed repeatedly (pre-dose and 0.5, 1, 1.5, 2, 3, 4 and 6 h) using Visual Analogue Scales (VAS) for ‘‘under the influence’’, ‘‘good drug effect’’ and ‘‘bad drug effect’’. VAS peak values were analysed with repeated-measures ANOVA (dose as within-subject factor) followed by Tukey post hoc tests; significance was set at p < 0.05. Onset/offset and effect duration were estimated from model-predicted ‘‘under the influence’’ curves using a 25% of maximum individual-response threshold.

Pharmacokinetics: Plasma LSD concentrations were quantifiable at all doses and for all analysed samples. Complete PK parameter estimation was possible for differing subsets of participants: 13 subjects for the 5 µg dose, 18 for 10 µg, and 15 for 20 µg, reflecting some missing samples due to technical problems. Concentration–time profiles followed first-order elimination and were well described by a one-compartment model. Compartmental estimates of elimination half-life were 2.5, 2.7 and 2.9 h for the 5, 10 and 20 µg doses, respectively; non-compartmental estimates reported in the discussion were slightly longer (approximately 3.0, 3.3 and 3.6 h). LSD concentrations increased in a dose-proportional manner across the 5–20 µg range. Between-subject variability in Cmax was relatively low (coefficient of variation approximately 30–36%). Subjective effects and PK–PD: The 5 µg dose produced no significant acute subjective effects versus placebo. The 10 µg dose produced statistically significant increases in VAS ‘‘under the influence’’ and ‘‘good drug effect’’ (both p < 0.05). At 10 µg, model-derived timings for the ‘‘under the influence’’ response averaged onset at 1.1 h, peak at 2.5 h and offset at 5.1 h, giving a modelled effect duration of about 4.0 h. The 20 µg dose produced significant increases in ‘‘under the influence’’, ‘‘good drug effect’’ and ‘‘bad drug effect’’ (all p < 0.001). PK–PD modelling yielded EC50 estimates indicating that concentrations producing half-maximal ‘‘good drug effects’’ were lower than those for ‘‘bad drug effects’’ (example values for the 10 µg condition: mean EC50 for good effects 0.86 ± 0.7, for bad effects 1.6 ± 0.8; units correspond to the modelled effect-site concentration). Overall, subjective effects closely paralleled plasma concentrations within individuals as reflected by the good model fit.

Zhang and colleagues interpret these data as the first comprehensive characterisation of LSD pharmacokinetics at very low single doses, enabled by a sensitive assay and frequent early sampling. The study found dose-proportional increases in plasma concentrations across 5–20 µg, first-order elimination kinetics, and short terminal half-lives of roughly 3 h, consistent with prior high-dose LSD studies and inconsistent with an earlier report of a markedly longer terminal half-life. The authors note that such a short half-life argues against bodily accumulation with repeated low-dose regimens, even with daily dosing. Pharmacodynamically, subjective effects tracked plasma concentrations closely, and the PK–PD modelling showed lower EC50 values for positive (‘‘good drug’’) effects than for negative (‘‘bad drug’’) effects. A threshold for perceivable subjective effects was identified at about 10 µg LSD base; doses below this (for example 5 µg) were effectively sub-perceptual in the present healthy young sample. The 20 µg dose produced small but measurable psychedelic-like effects in some participants, so it is characterised as a very low to low psychedelic dose rather than truly sub-perceptual. The investigators highlight strengths including the within-subject crossover design, double-blinding, use of a validated low-LLOQ assay (2.5 pg/ml), and PK–PD modelling across multiple subjective endpoints. They also acknowledge limitations: plasma samples were not available from every participant for every dose, resulting in different subject numbers across PK analyses; the trial focused on psychological measures rather than broader safety or physiological endpoints; and the standardised pharmaceutical formulation used here may not reflect the pharmacokinetics of recreational products (commonly LSD tartrate on blotter paper). The authors call for confirmatory pharmacokinetic studies with more complete sampling and note that responses could differ in clinical populations or in less tightly monitored real-world settings. Finally, they recommend use of well-characterised LSD formulations in future research to ensure consistent exposure and facilitate comparisons between studies.