Genetic influence of CYP2D6 on pharmacokinetics and acute subjective effects of LSD in a pooled analysis

Earlier research indicates that LSD is metabolised in the liver and that several cytochrome P450 (CYP) enzymes—particularly CYP2D6, CYP1A2, CYP2C9, CYP2C19 and CYP3A4—can contribute to its biotransformation in vitro, but in vivo pharmacogenetic data have been lacking. CYP2D6 shows common functional genetic polymorphisms that produce phenotypes ranging from poor metabolizers (PMs) to ultra-rapid metabolizers, and previous pharmacogenetic work with other psychoactive drugs (for example MDMA) suggests that CYP genotype can alter drug exposure and acute effects. Because acute subjective effects of psychedelics are closely linked to plasma concentration–time profiles and may predict therapeutic outcomes, understanding genetic influences on LSD pharmacokinetics and acute response is clinically relevant. Vizeli and colleagues set out to test whether common functional variants of several CYP genes influence LSD pharmacokinetics and its acute subjective effects in humans. Guided by in vitro findings, the primary hypothesis was that individuals genetically defined as CYP2D6 poor metabolizers would exhibit higher LSD concentrations and stronger acute effects compared with carriers of functional CYP2D6 alleles. The study pooled data from four randomized, double-blind, placebo-controlled Phase I crossover trials to address this question in a sample of healthy adults.

Design and setting: The paper reports a pooled secondary analysis of four Phase I studies run at the same laboratory, each using randomized, double-blind, placebo-controlled, crossover protocols. The four parent studies (registered at ClinicalTrials.gov) provided comparable controlled test sessions in a hospital research ward; washout intervals were 7–10 days depending on the study. Participants: Across the four studies, 85 healthy subjects of European descent aged 25–60 years were enrolled; after withdrawals and two subjects declining genotyping the final pooled sample comprised 81 participants (41 women), mean age 30 ± 8 years and mean weight 70 ± 12 kg. Exclusion criteria across the parent studies included history of psychiatric disorder, significant physical illness, heavy smoking (>10 cigarettes/day), concurrent medications likely to interfere with study drugs, and substantial recent illicit drug use. A minority (22 participants) had limited prior hallucinogen experience. Interventions and dosing: LSD was administered orally in doses typical for experimental or therapeutic settings. Study 1 used a single 200 µg dose, Studies 2 and 3 used 100 µg, and Study 4 included 25, 50, 100 and 200 µg doses (Study 4 also included a 200 µg + 40 mg ketanserin condition used only in pharmacokinetic analyses). For the pooled analysis the investigators used the available dose data per subject; for modelling and group comparisons of the 100 µg condition, only the 100 µg data were included. LSD was given after a standardised small breakfast; subjects were monitored and remained under supervision during acute effects. Pharmacokinetic and physiological measurements: Plasma LSD and the major metabolite 2-oxo-3-hydroxy-LSD (O-H-LSD) were assayed and pharmacokinetic parameters calculated using non-compartmental analysis in Phoenix WinNonlin 6.4, with AUC values computed to last measured concentration (AUC10) and extrapolated to infinity (AUC∞). Peak effect (Emax) and area under the time–effect curve (AUEC) were derived for subjective measures. A one-compartment model with first-order input and elimination was applied to illustrate concentration–time profiles and to compare CYP2D6 groups for the 100 µg dose. Genotyping and phenotype assignment: Genomic DNA from whole blood was genotyped for common SNPs and structural variants across CYP2D6, CYP1A2, CYP2C9, CYP2C19 and CYP2B6 using TaqMan assays and CNV analysis for CYP2D6 deletions/duplications. Activity scores for CYP2D6 were assigned according to established guidelines and subjects were dichotomised for primary analyses into non-functional CYP2D6 (PMs; activity score = 0) versus functional CYP2D6 (activity score > 0). CYP1A2 inducibility was combined with smoking status (>5 cigarettes/day) to approximate activity; no CYP2C19 PMs were identified in the sample. Outcomes and statistics: Acute subjective effects were measured with validated psychometric scales including the 5D-ASC and visual analogue scales (VAS); physiological variables (blood pressure, heart rate, temperature) were repeatedly recorded. Genotype effects on pharmacokinetic parameters and Δ(LSD–placebo) effects were tested using one-way ANOVA with genotype as between-group factor. To address potential non-normality and outliers, complementary non-parametric tests (Wilcoxon signed-rank and Kruskal–Wallis) were applied. Analyses reported both nominal values and study-wise z-scores to adjust for between-study dose differences. The significance threshold was p < 0.05; pharmacokinetic p-values were not corrected for multiple testing where hypotheses were prespecified (for example, CYP2D6).

Sample and general effects: The pooled dataset comprised 81 genotyped participants. LSD produced robust acute subjective effects across psychometric scales and produced moderate increases in blood pressure, heart rate and body temperature relative to placebo. Neither sex nor body weight significantly influenced LSD exposure or acute effects in these analyses. CYP2D6 effects on pharmacokinetics: Genetically defined CYP2D6 functionality significantly affected LSD pharmacokinetics. Subjects classified as CYP2D6 poor metabolizers (activity score = 0) exhibited substantially higher plasma LSD exposure, reflected in larger AUC∞ and AUC10 values compared with carriers of functional CYP2D6 alleles. CYP2D6 PMs also had longer elimination half-lives (T1/2), consistent with slower LSD clearance, whereas peak concentration (Cmax) differences were not statistically significant. In compartmental modelling of the 100 µg dose, mean (±SD) AUC∞ for PMs versus functional subjects was 24,169 ± 13,112 versus 13,819 ± 6,281 pg·ml−1·h (F1,79 = 13.8, p < 0.001), and Cmax was 2,369 ± 891 versus 2,061 ± 999 pg·ml−1 (F1,79 = 0.62, p = 0.43). Across activity-score groups there was a dose–response relationship with lower CYP2D6 activity associated with higher LSD exposure. The authors report that CYP2D6 PMs showed approximately 75% greater total drug exposure (AUC∞) than functional CYP2D6 individuals, while mean peak concentration was only about 15% higher and not significant. Metabolite concentrations: O-H-LSD AUC∞ values were also larger in CYP2D6 PMs, paralleling parent drug effects. The investigators interpret this as indicating that conversion to O-H-LSD occurred independently of CYP2D6, while CYP2D6 contributes importantly to other degradation pathways (for example N-demethylation to nor-LSD). CYP2D6 effects on subjective outcomes: The altered pharmacokinetic profile in CYP2D6 PMs was mirrored by differences in subjective response. Poor metabolizers experienced a substantially longer duration of acute subjective effects and significantly greater alterations of mind on the 5D-ASC total score, and on subscales reflecting anxious/ego-disintegration type effects (AED: disembodiment, impaired control and cognition, anxiety) and visual/reperceptual phenomena (VR: complex and elementary imagery, changed meaning of percepts). Peak VAS effects were not different between genotypes, consistent with similar Cmax values. There were no genotype-dependent differences in autonomic measures. Other CYPs: Genetic polymorphisms of CYP1A2, CYP2B6, CYP2C19 and CYP2C9 showed no relevant effects on LSD pharmacokinetics, subjective outcomes or autonomic responses in this sample.

Vizeli and colleagues interpret their findings as the first in vivo pharmacogenetic evidence that CYP2D6 polymorphism meaningfully alters LSD metabolism and, in part, its acute subjective effects in humans. The data indicate that CYP2D6 contributes to LSD degradation—consistent with in vitro reports implicating CYP2D6 in N-demethylation to nor-LSD—because CYP2D6 poor metabolizers had prolonged half-lives and greater total LSD exposure without a substantial increase in peak concentration. The concurrent increase in O-H-LSD in PMs suggests that O-H-LSD formation proceeds independently of CYP2D6 and that the observed AUC effects reflect reduced clearance via other CYP2D6-mediated pathways. The investigators discuss clinical implications, noting that co-administration of CYP2D6 inhibitors (for example some SSRIs such as fluoxetine or paroxetine) could similarly increase LSD exposure and prolong effects, and that allowing time for enzyme recovery or dose reduction might be prudent when combining treatments. They also contrast LSD with MDMA: for MDMA, CYP2D6 genotype effects are limited by MDMA's autoinhibition of CYP2D6 and are most evident early after dosing, whereas for LSD CYP2D6 genotype appears to affect elimination and overall exposure more than early absorption or peak. On subjective effects, CYP2D6 PMs reported higher scores on AED and VR subscales but not on oceanic boundlessness (OB); because certain qualities of the acute psychedelic experience have been linked to therapeutic outcomes in prior studies, the authors note that greater AED and VR but not OB could potentially result in a more challenging acute experience and may not predict improved therapeutic benefit—an issue that requires further study. As a practical suggestion, the authors propose that CYP2D6 genotyping or phenotyping might inform dosing in LSD-assisted therapy, and they estimate that PMs might benefit from approximately 50% lower doses than functional metabolizers, while stressing that this requires validation. Limitations acknowledged by the authors include the relatively small sample size despite pooling the largest available controlled dataset, which may limit power to detect smaller effects for other CYPs and preclude assessment of rarer variants (for example CYP3A4 polymorphisms). They also note the possibility of type I errors despite prespecified hypotheses. Strengths cited include the placebo-controlled crossover designs of the parent studies, use of validated psychometric instruments, complementary non-parametric statistical checks, and z-transformation to account for between-study dose differences. Finally, the authors call for confirmatory drug–drug interaction studies with selective CYP inducers/inhibitors and for further work to evaluate the utility of pharmacogenetic testing before LSD-assisted psychotherapy.