Acute Effects and Pharmacokinetics of LSD after Paroxetine or Placebo Pre-Administration in a Randomized, Double-Blind, Cross-Over Phase I Trial

Classic psychedelics such as LSD and psilocybin act principally via serotonin 5-HT2A receptor activation and are under investigation as treatments for depressive and anxiety disorders. Because many patients eligible for psychedelic-assisted therapies are already taking selective serotonin reuptake inhibitors (SSRIs), concerns exist that SSRIs might modulate the acute effects of psychedelics via effects on 5-HT2A receptor number or function; case reports and survey data have suggested attenuation of psychedelic responses during SSRI treatment. Prior controlled work from this group showed reduced acute response to psilocybin after escitalopram, and clinical trials commonly discontinue SSRIs before psychedelic dosing despite withdrawal and relapse risks associated with discontinuation. Becker and colleagues designed the present trial to test whether six weeks of paroxetine (a strong CYP2D6 inhibitor) versus placebo would alter the acute subjective, adverse and autonomic responses to a single 100 μg dose of LSD in healthy volunteers. The study also aimed to characterise pharmacokinetic interactions and the role of CYP2D6 genotype in LSD metabolism, and to test the primary hypotheses that overall mind-altering effects (3D-OAV/5D-ASC total score) and peak "good drug effect" VAS ratings would be similar after paroxetine and placebo, while "bad drug effect" and "anxiety" would be reduced after paroxetine. The paroxetine manipulation additionally allowed investigation of whether increased LSD exposure via CYP2D6 inhibition would offset any SSRI-related pharmacodynamic attenuation.

This was a randomized, double-blind, placebo-controlled, cross-over Phase I trial with two 26-h experimental sessions per participant. Healthy volunteers aged 25–65 were recruited and screened; key exclusions included pregnancy, first-degree family history of psychosis, current or past major psychiatric disorder, certain medication use, significant physical illness, heavy tobacco or alcohol use, recent illicit drug use, and lifetime psychedelic use >20 times. Participants self-administered either paroxetine (10 mg daily for 7 days then 20 mg daily for 35 days) or matching placebo for 42 days before each LSD session, with the order randomised and counterbalanced. The last capsule was taken on site 1 h before LSD. Compliance was checked by pill counts and measurement of plasma paroxetine; sessions were separated by at least a 2-day washout. No placebo for LSD was used. LSD D-tartrate (146 μg, equivalent to 100 μg LSD base) was given as a drinking solution at 09:30 on each test day. Subjective effects were assessed with 17 visual analogue scales (VASs) repeatedly up to 24 h, with the VAS item "good drug effect" predefined as a primary pleasant-effects endpoint. The 5D-ASC, including the 3D-OAV total score, served as the primary endpoint for overall alterations of mind; additional measures included the States of Consciousness Questionnaire (PES48 and MEQ30) and the Adjective Mood Rating Scale (AMRS). Adverse effects were recorded with the List of Complaints and by AE reporting; nausea was measured both on the LC and a dedicated VAS. Autonomic measures included repeated blood pressure, heart rate and tympanic temperature readings and ECG before and 3.5 h after LSD. Pharmacokinetic sampling occurred at the same time points as the VAS measures. Plasma LSD and O-H-LSD concentrations were quantified by LC-MS/MS (LLOQ 10 pg/mL); paroxetine was assayed with an adapted LC-MS/MS method (LLOQ 1 ng/mL). Non-compartmental analysis provided Cmax, Tmax, half-life and AUC parameters; PK-PD modelling was also performed. CYP2D6 genotyping was conducted to classify participants as normal, intermediate or poor metabolizers, and whole-blood gene expression of HTR2A, SLC6A4 and BDNF was measured. Statistical analyses used paired two-sided t-tests for repeated measures with p < 0.05 as the significance threshold. For PK interactions, geometric mean ratios (GMRs) with 90% confidence intervals were computed for AUC∞ and Cmax according to FDA guidance, with a default no-effect range of 0.8–1.25.

Of 37 screened individuals, 27 entered the study; after three dropouts and one noncompliant participant with undetectable paroxetine, 23 participants contributed data (12 male, 11 female; mean age 31 ± 8 years, range 25–55; mean weight 68 ± 11 kg). Ten participants had prior psychedelic use (mean 6 ± 5 occasions). On subjective measures, paroxetine significantly reduced LSD-induced single-item VAS ratings of "bad drug effect" and "anxiety," while not affecting "good drug effect" or the 3D-OAV total score on the 5D-ASC. The duration of overall LSD effect was similar between conditions (paroxetine 9.1 ± 2.3 h; placebo 9.6 ± 2.4 h). No significant differences emerged across 5D-ASC dimensions or factors, though there was a non-significant trend toward more pleasant effects (e.g. "blissful state") and higher PES48 ratings after paroxetine. AMRS mood changes did not differ between conditions. Adverse-event reporting showed similar acute (0–12 h) and subacute (12–24 h) profiles across conditions, with a tendency toward fewer total adverse effects after paroxetine (acute: placebo 181 vs paroxetine 158; subacute: placebo 52 vs paroxetine 27). Headache was most frequent (placebo: 13 participants; paroxetine: 16). Nausea was reported by 13 participants after placebo and 6 after paroxetine; peak and overall nausea VAS scores were significantly lower with paroxetine. Two participants experienced prolonged nausea with emesis in the placebo condition only; domperidone was given once without improvement. Three administrations of paracetamol occurred in each condition for headaches. No severe adverse events were observed. Paroxetine attenuated LSD-induced heart rate increases but did not alter blood pressure effects. QTc interval did not increase at peak effect (3.5 h) compared with baseline in either condition and no QTc exceeded 450 ms. Pharmacokinetics showed that paroxetine increased LSD exposure: AUC∞ GMR 1.47 (90% CI 1.29–1.68), Cmax GMR 1.41 (1.24–1.60), and half-life ratio 1.24 (1.17–1.31), exceeding the default no-effect boundaries. Geometric mean paroxetine concentration before the final dose was 17.0 ng/mL. CYP2D6 genotyping identified 11 normal metabolizers (NM), 9 intermediate metabolizers (IM) and 3 poor metabolizers (PM). AUC∞ and Cmax were highest in PMs and decreased with increasing CYP2D6 activity. Paroxetine had little effect on LSD exposure in PMs, whereas the magnitude of paroxetine/placebo ratios for AUC∞ and Cmax was greater in participants with higher CYP2D6 activity, supporting a CYP2D6-mediated contribution to LSD metabolism; O-H-LSD showed a similar pattern. Whole-blood expression of HTR2A, SLC6A4 and BDNF did not differ between paroxetine and placebo conditions. Blinding was imperfect: participants correctly identified their assigned condition in 83% of sessions, commonly due to paroxetine side effects. End-of-study retrospective VAS ratings of maximal LSD effects did not differ significantly between conditions, although 70% of participants judged LSD after paroxetine to be more pleasant and 61% judged LSD after placebo to be more valuable; these differences were not statistically significant.

Becker and colleagues interpret the findings as confirming that six weeks of paroxetine pre-administration produced equivalent overall intensity and pleasant subjective effects of LSD while reducing unpleasant acute responses. Specifically, paroxetine decreased peak "bad drug effect," anxiety and nausea without worsening cardiovascular safety markers or prolonging QTc. These results align with prior work from the group showing escitalopram reduced challenging subjective effects of psilocybin but preserved pleasant effects. Pharmacokinetic data indicated that paroxetine, a strong CYP2D6 inhibitor, raised LSD peak concentrations and overall exposure by approximately 41% and 47% respectively, and prolonged LSD half-life. The CYP2D6 genotype analysis supported a role for CYP2D6 in LSD metabolism: poor metabolizers had the highest exposures and showed little modulation by paroxetine, while the paroxetine effect was greater in participants with higher baseline CYP2D6 activity. The authors note a pharmacodynamic interaction — a higher EC50 for LSD after paroxetine consistent with reduced potency — but that the CYP2D6-mediated pharmacokinetic increase in LSD exposure likely compensated for decreased pharmacodynamic responsiveness, resulting in similar overall subjective intensity. The study did not detect changes in peripheral gene expression of HTR2A, SLC6A4 or BDNF after paroxetine, and the authors caution that peripheral measures may not reflect central receptor changes. They acknowledge limits to generalisability: participants were healthy and relatively young, the six-week run-in is shorter than typical clinical SSRI use, and results may not extrapolate to chronic SSRI-treated patients. The discussion highlights practical implications raised by the findings: maintaining SSRIs during psychedelic-assisted therapy could avoid discontinuation risks and may even reduce acute anxiety and nausea, but controlled patient trials are needed to determine whether therapeutic efficacy is preserved when antidepressants are continued. Strengths cited include the randomised, cross-over design, clinically relevant dosing and comprehensive assessment; limitations include the sample demographics, the limited run-in duration relative to long-term SSRI exposure and the need to study interactions with other antidepressants and psychedelics. The authors call for Phase II trials in patients to compare therapeutic outcomes with and without concomitant antidepressant pharmacotherapy.

Paroxetine pre-administration did not alter the acute pleasant subjective effects of a 100 μg LSD dose but significantly reduced "bad drug effect," anxiety and nausea. As a strong CYP2D6 inhibitor, paroxetine increased LSD plasma concentrations, supporting CYP2D6 involvement in LSD metabolism.