LSD-assisted therapy in patients with anxiety: open-label prospective 12-month follow-up

Anxiety disorders are common, often chronic, and current treatments can be slow to produce benefit, leaving a need for new therapeutic approaches. Early clinical studies of classic psychedelics such as LSD and psilocybin have reported rapid and sometimes sustained reductions in anxiety and depressive symptoms after single-dose treatment, and there is emerging evidence that psychedelic experiences can induce changes in personality traits relevant to mood and anxiety disorders. However, the durability of these effects over longer follow-up intervals remains uncertain, and long-term safety data are limited. Holze and colleagues report a planned 12-month follow-up of participants from a previously published double-blind, placebo-controlled, crossover trial of LSD-assisted therapy in people with anxiety disorders, with or without life-threatening illness (LTI). The study aimed to assess whether reductions in anxiety and comorbid depression persisted one year after the end-of-study visit, and to explore enduring changes in general psychiatric symptoms, personality traits and self‑reported positive or negative long-term effects attributed to the psychedelic experience. The authors stated an initial hypothesis that therapeutic effects would not be sustained to the 12-month follow-up, and also examined correlations between acute subjective experiences during LSD sessions and long-term outcomes.

This report is an a priori-planned long-term (12-month) follow-up of an investigator‑initiated, two-centre, double-blind, placebo-controlled, random-order crossover trial. Each participant completed two 24-week treatment periods; in each period they received two oral sessions of either LSD (200 μg per session) or matched placebo, with sessions in a period separated by about 6 weeks. The order of LSD versus placebo periods was randomised and counterbalanced. Follow-up questionnaires were mailed to all participants 12 months after the end-of-study visit. Because of the crossover design and scheduling, the follow-up assessments occurred at different intervals from the participants' last LSD dose: for those who received LSD in the first period the follow-up corresponded to 94 weeks after their last LSD treatment, and for those who received LSD in the second period it corresponded to 68 weeks after the last LSD treatment. Participants were adults with anxiety disorders, recruited to include people with anxiety associated with LTI (defined as a severe somatic disease such as cancer or another potentially fatal illness) and people without LTI. Inclusion required a DSM-IV anxiety diagnosis for participants without LTI; for participants with LTI a DSM-IV anxiety diagnosis or an elevated STAI score (≥40) sufficed. All participants gave written informed consent and were assigned to a single investigator/therapist for the duration of the trial. Active study drug was GMP-produced LSD free base (>99% purity) administered as an oral ethanol solution delivering 100 μg/mL (two mL per active session to yield 200 μg), and placebo consisted of identical ethanol-only units. Primary and secondary outcomes were assessed by self-report questionnaires. The primary outcome for the original trial was the Spielberger State-Trait Anxiety Inventory–Global (STAI-G) with a primary endpoint at 16 weeks after the last treatment session; this long-term report compares follow-up scores with baseline. Secondary measures included STAI State and Trait subscales (STAI-S, STAI-T), Beck Depression Inventory (BDI), and the SCL-90-R for general psychiatric symptoms. Persisting effects were assessed using the 143‑item Persisting Effects Questionnaire (PEQ) and personality traits were measured with the 60‑item NEO Five-Factor Inventory (NEO‑FFI) at screening and at 1-year follow-up. Acute subjective drug effects during sessions had been measured previously with the 5D-ASC and the MEQ30. Adverse event queries at follow-up included items on flashbacks and hallucinogen persisting perception disorder (HPPD). Analyses used linear mixed-effects models with time points as fixed factors and participants as random factors; baseline values for the period were included as covariates for therapeutic outcomes. Tukey post hoc pairwise comparisons of least square means were used to compare baseline and follow-up. Exploratory Pearson correlations assessed associations between acute-session measures (e.g. MEQ30, 5D-ASC) and long-term outcomes. The analysis plan and sample size calculation were described in the original protocol; the extracted text reports that analyses were conducted in R with lme4/lmerTest and a significance threshold of P < 0.05. Follow-up questionnaires were sent by post and responses included those from study completers and some dropouts.

Enrolment ran from June 2017 to February 2021, with the last patient visit in December 2021 and follow-up data collection completed by January 2023. Of the original cohort, follow-up questionnaires were returned by 35 individuals (15 with LTI, 20 without LTI). For STAI measures 33 participants had usable data and for other measures 34 participants were available. For change-from-baseline analyses per treatment group, 39 participants (20 in the LSD‑first group and 19 in the placebo‑first group) were eligible. Six participants subsequently received additional psychedelic therapy outside the trial (within a Swiss limited‑use programme), all with further LSD sessions. Participants reported sustained reductions in anxiety, depression and general psychiatric symptoms at 12 months after the end-of-study visit compared with baseline. For the primary measure STAI-G, least square mean change from baseline to follow-up was −21.6 (95% CI −32.7 to −10.4) for the LSD‑first group (corresponding to week 102 in the study timeline) and −16.5 (−26.2 to −6.8) for the placebo‑first group (week 76); these changes were statistically significant (P < 0.001 and P < 0.05, respectively) and associated with large effect sizes (Cohen's d = 1.04 and 1.02). The LSD‑first group showed further improvement from the primary 16‑week endpoint to the 12‑month follow-up (least square mean change −9.7, 95% CI −15.5 to −4.0), whereas the placebo‑first group showed no further change (1.2, −4.2 to 6.6). Post hoc tests indicated no significant differences in follow-up change by LTI status, by whether participants had additional psychedelic therapy in the limited‑use programme, or by a formal diagnosis of generalised anxiety disorder. Clinical remission at follow-up was reported for 11 participants (33%) with respect to anxiety symptoms and for 17 participants (49%) with respect to depression symptoms, as presented by the authors. On the PEQ, participants gave high ratings for positive long-term effects (positive attitudes about life and self, positive mood, altruistic/positive social effects and positive behavioural changes, each >47% of the maximum score) while negative long-term effect ratings remained low (<10% of maximum). On average, participants rated the psychedelic experience as among the ten most personally meaningful experiences of their lives and as ‘‘very much’’ spiritually meaningful, and reported positive influence on well‑being or life satisfaction. The LSD‑first group showed nominally higher positive PEQ ratings and nominally lower negative item ratings than the placebo‑first group. Personality measures changed between screening and the 1‑year follow-up: neuroticism decreased significantly (P < 0.0001) and extraversion increased significantly (P = 0.002) on the NEO‑FFI; other personality domains did not show significant change in the extracted text. Correlational analyses showed that larger reductions in STAI‑G (ΔSTAI‑G) were significantly associated with greater mystical‑type acute experiences (MEQ30 total score; r = −0.342, P = 0.048) and showed a near‑significant association with oceanic boundlessness on the 5D‑ASC (r = −0.332, P = 0.055). Acute anxious ego‑dissolution did not correlate with long-term anxiety change. Several positive PEQ subscales (positive attitudes about life/self, positive mood changes, altruistic/social effects) correlated significantly with MEQ30 and with oceanic boundlessness on the 5D‑ASC; negative acute effects showed no association with PEQ outcomes. No additional adverse events were reported during the follow-up period, and no clinically relevant flashbacks or HPPD occurred. Written feedback from 18 participants was predominantly positive, with comments on therapeutic potential, setting, therapist interaction and dosing; two participants gave predominantly negative feedback.

Holze and colleagues interpret the findings as evidence of potential long‑lasting benefits from LSD‑assisted therapy in people with anxiety symptoms, with sustained reductions in global anxiety (STAI‑G), comorbid depressive symptoms (BDI) and general psychiatric symptoms (SCL-90‑R) at 12 months post–end of study. Participants also attributed durable positive life changes to their psychedelic experiences, and the observed decreases in neuroticism alongside increases in extraversion are presented as changes consistent with improved clinical status, given the linkage of high neuroticism and low extraversion to mood and anxiety disorders. The authors note that these results contradict their a priori hypothesis that effects would not persist to 12 months. They place the findings alongside previous long‑term follow-up studies of psilocybin and other psychedelics in clinical and healthy samples, highlighting similarities in effect sizes and in self‑reported meaningfulness and positive life change. Holze and colleagues discuss the pattern of personality change in the context of prior reports that openness often increases after psychedelics; they also point out that participants in this trial had high baseline openness relative to population norms, which may have influenced recruitment and response. Possible explanations for group differences—particularly the apparently greater benefit in the LSD‑first group—are considered but remain unresolved. The authors highlight baseline symptom differences between groups and raise the possibility of expectancy effects or that two placebo sessions in the second period may have functioned as additional integration opportunities for those who had earlier received LSD. They recommend that future trials include validated expectancy measures. Regarding mechanisms, correlations between acute mystical‑type experiences and short‑term therapeutic response were stronger in the primary report than at long‑term follow-up in this analysis, suggesting that other factors—such as neuroplastic changes, psychotherapeutic integration, environmental influences or additional pharmacological/extrapharmacological variables—may contribute to enduring outcomes. Safety findings are emphasised: no new long‑term adverse events, clinically relevant flashbacks or HPPD were observed during follow-up. Strengths of the study cited by the authors include inclusion of participants with and without LTI, long within‑subject observation spanning approximately two years per participant, and a comparatively large sample for a psychedelic long‑term follow-up. Limitations acknowledged are the absence of a concurrent control group for long‑term outcomes, reliance on self‑report instruments rather than interview‑based assessments at follow-up, lack of an expectancy measure, and incomplete information about therapies participants may have received between the end‑of‑study visit and the 12‑month follow-up. The authors also note that all participants had received LSD within the crossover design, so long‑term comparisons are made against pre‑LSD baseline within subjects and other contributing factors cannot be excluded. In sum, one year after the last study visit participants continued to report reduced anxiety and depression, minimal lasting negative effects, improved well‑being and shifts in neuroticism and extraversion.