Single Treatment With MM120 (Lysergide) in Generalized Anxiety Disorder

Study Design and Participants

In this phase 2b, multicenter, double-blind, placebo-controlled trial, participants with GAD were randomized equally to receive a single (freebase equivalent) treatment dose with 25 µg, 50 µg, 100 µg, or 200 µg of MM120 or placebo without stratification (Figure 1). The details of dose selection, dosing procedures, and study methods appear in the trial protocol (Supplement 1), in the statistical analysis plan (Supplement 2), and in the eMethods and eTable 1 in Supplement 3.

The eligibility criteria included age of 18 to 74 years, having a Diagnostic and Statistical Manual of Mental Disorders (Fifth Edition) primary diagnosis of GAD based on investigator assessment and confirmed by a Mini-International Neuropsychiatric Interview, and having moderate to severe anxiety (Hamilton Anxiety Rating Scale [HAM-A] score of ≥20) at both screening and at baseline. Participants undergoing psychotherapy at the time of screening could enroll but could not discontinue or modify session frequency during the study. Participants not receiving psychotherapy at screening could not initiate it during the trial. The exclusion criteria included contraindicated medical or psychiatric conditions and inability or unwillingness to discontinue therapeutics with potential drug-drug interaction or anxiolytic effects.

Eligibility was evaluated at screening and baseline, with final confirmation on dosing day via a urine drug screening, urine pregnancy test, and collection of vital signs. Disclosure or detection of the use of prohibited substances, including tetrahydrocannabinol, was cause for screening failure or discontinuation as was the use of any psychedelic within the past 2 years or more than 10 times within the past 10 years.

All participants provided written informed consent that included details of the known effects of MM120. Participants self-reported race and ethnicity using categories from the 2020 US Census (multiple selections were allowed). The study was approved by a central institutional review board and local institutional review boards were engaged for the sites when required. The study was conducted at 22 outpatient psychiatric research sites across the US (eTable 2 in Supplement 3). The Consolidated Standards of Reporting Trials (CONSORT) reporting guideline was used.

Procedures

The study included 3 phases: (1) eligibility assessment (up to 35 days), (2) randomization (1 day), and (3) follow-up (12 weeks) (eFigure 1 in Supplement 3). Participants taking mood or anxiety medications at screening were required to discontinue them under the supervision of site physicians. Before the baseline visit, a medication washout period of 5 half-lives or greater was required. During the baseline visit, the dosing session monitors provided information to the participants on session logistics, staff interactions, and the acute effects of MM120.

On the dosing day, participants received a single oral dose of MM120 or placebo. Each participant was continuously monitored on-site for safety and comfort by the same 2 dosing session monitors who conducted the participant’s educational session during the baseline visit. Participant dosing sessions were conducted individually in a private aesthetically pleasant room. The dosing sessions were not recorded or conducted in a group setting. Participants were offered standardized music and eyeshades and they could lie down, move freely around the room, read, write, or draw.

The dosing session monitors assisted participants with functional needs (such as eating or being escorted to the restroom) upon participant request. Vital signs were measured throughout the dosing session and a physician was on-site in case comfort medications were needed. The dosing session monitors received study-specific training regarding how to ensure the comfort of the participants and safety management. Engaging in psychotherapy with the participant was explicitly prohibited by the trial protocol (Supplement 1).

Starting 8 hours after the dose, participants were assessed hourly to characterize the resolution of the drug effects. Continuous monitoring was required for a minimum of 12 hours. Dosing session monitors met with participants at follow-up visits to assess safety and answer any questions about the dosing day (these visits occurred on day 2, at week 1, and at week 2). Efficacy and safety evaluations were conducted by site staff during follow-up visits (on day 2 and at weeks 1, 2, 4, 8, and 12).

Outcomes

Trained, certified raters assessed all outcome measures. To mitigate bias, the dosing session monitors were prohibited from conducting the efficacy assessments. The HAM-A and the Montgomery-Åsberg Depression Rating Scale (MADRS) scores were assessed by independent central raters via telephone; the raters were blinded to the trial protocol, treatment allocation, and study visit.

The primary outcome was a dose-response relationship assessed using the multiple comparison procedure modeling (MCP-Mod) method for change in HAM-A score from baseline to week 4 (score range, 0-56; higher scores indicate greater severity; ≤7 indicates no or minimal anxiety; 8-14, mild anxiety; 15-23, moderate anxiety; and ≥24, severe anxiety²⁶). The HAM-A score was assessed by independent central raters using the Structured Interview Guide for the Hamilton Anxiety Rating Scale. The key secondary outcome was a dose-response relationship^27,28 assessed using the MCP-Mod method for change in HAM-A score at 8 weeks.

Additional secondary outcomes included (1) change in HAM-A score at weeks 1, 2, 4, 8, and 12; (2) change in depressive symptom score (measured using the MADRS score; range, 0-60; higher scores indicate greater severity)²⁹ at weeks 1, 2, 4, 8, and 12; and (3) change in clinician-rated illness severity (measured using the Clinical Global Impression–Severity [CGI-S] score; range, 1-7; higher scores indicate more severe illness) at day 2 and at weeks 1, 2, 4, 8, and 12. The rates of response (≥50% change [reduction] in HAM-A score from baseline) and remission (HAM-A score ≤7) were also explored. The effects of MM120 on functional disability, quality of life, and sexual function were assessed throughout the trial using the Sheehan Disability Scale, the EuroQol 5-dimension 5-level questionnaire, the Pittsburgh Sleep Quality Index, and the Arizona Sexual Experiences Questionnaire. An entire list of prespecified outcomes appears in Supplement 1.

Blinding of the participants and raters regarding treatment allocation (active drug, placebo, or indiscernible) was assessed using a 5-point questionnaire (range, 1 [positive they received drug] to 5 [positive they received placebo]).^30,31 Safety and tolerability were evaluated through adverse event (AE) monitoring, physical and neuropsychiatric examinations, vital signs, electrocardiographic monitoring, and laboratory tests. Suicidal ideation and behavior were assessed at each visit using the Columbia–Suicide Severity Rating Scale (C-SSRS).³²

Statistical Analysis

Efficacy was analyzed in the full analysis set, which was composed of all randomized participants with a HAM-A measurement at baseline and at least 1 HAM-A measurement after baseline. Safety was evaluated in all participants who were randomized.

The primary outcome and the key secondary outcome were evaluated using the MCP-Mod approach at the 5% significance level, which integrates hypothesis testing with dose-response modeling and is endorsed by regulatory agencies for dose selection.^33,34 The statistical significance of the candidate set of dose-response models was assessed separately at week 4 (for the primary outcome) and at week 8 (for the key secondary outcome).

Six candidate dose-response models were prespecified to represent a range of plausible dose-response shapes (model [model parameters]; sigmoid Emax [10,1], sigmoid Emax [100,5], sigmoid Emax [100,10], sigmoid Emax [150,10], Emax [100], and linear). The objective of the primary outcome and the key secondary outcome was to statistically detect a dose-response relationship, which was assessed under strict type I error control at the 5% level using a hierarchical gatekeeping testing procedure. Missing primary end point data were imputed using multiple imputations with 20 iterations.

Observations missing due to prohibited medication use were assumed to be missing not at random and imputed using placebo-based imputation. All other missing observations were imputed under the assumption of missing at random. Both imputation models included all time points after baseline and were implemented using the fully conditional specification framework with predictive mean matching. Baseline was included as a covariate and treatment group was incorporated either by stratification or as a model variable.

To achieve a mean power greater than 87% for the MCP-Mod approach analysis, 200 participants were planned to yield at least 180 evaluable end points after any dropouts. A placebo-adjusted difference of 2.5 points for change in HAM-A score was used as the threshold (minimal clinically important difference of 2.5 points) for determining the minimally efficacious dose.³⁵ Statistical significance was set at an α level of .05. The estimated dose-response curve was derived using an Akaike information criterion–weighted average of the top 3 dose-response models.

Change in HAM-A score from baseline was analyzed using an analysis of covariance model adjusted for the baseline values and treatment group without imputation (observed cases only). The analysis of covariance was summarized using least-squares means (with 95% CIs) and least-squares mean differences (with 95% CIs) vs placebo. Cohen d was calculated as the bias-corrected least-squares estimate of the treatment difference in changes from baseline to a given time point (from an analysis of covariance model), divided by the pooled SD of observed values at the same time point from a given active dose group and the placebo group.

Other secondary efficacy outcomes were analyzed using 2-sample t tests. For a given end point and time point, 4 t tests were used to compare changes from baseline for each MM120 dose group vs placebo. For the secondary outcomes focused on functional disability, the changes from baseline were analyzed descriptively. The categorical data were summarized and presented as counts and percentages without statistical testing.

All analyses were conducted using SAS version 9.4 (SAS Institute Inc) and the Dose Finding R package version 1.1-1 (R Foundation for Statistical Computing). Additional details appear in the statistical analysis plan (Supplement 2).

Participants

Between August 24, 2022, and August 30, 2023, a total of 554 participants were screened; of these, 198 were randomized and received a single (freebase equivalent) treatment dose with 25 µg (n = 39), 50 µg (n = 40), 100 µg (n = 40), or 200 µg (n = 40) of MM120 or placebo (n = 39) (Figure 1). Of the 198 participants randomized, 194 were included in the full analysis set (mean age, 41.3 [SD, 13.6] years; 56.7% were female; and 3.6% were Asian, 7.7% were Black or African American, and 83.0% were White) (Table 1). The full analysis set was composed of all randomized participants with a HAM-A measurement at baseline and at least 1 HAM-A measurement after baseline.

At screening, 84% of participants required tapering prohibited medications, 16% were treatment-naive, and 18% were receiving psychotherapy at the time of enrollment; psychotherapy was allowed to continue through the trial. Of the participants who required tapering prohibited medications, 43% were taking serotonin reuptake inhibitors, 34% were taking benzodiazepines, and 19% were taking antipsychotics or antiepileptics. Participants may have tapered multiple medications. The mean HAM-A scores at baseline were equivalent across groups (score range, 29.3-31.0).

Efficacy

A significant MM120 dose-response relationship was demonstrated at both week 4 (primary outcome) and week 8 (key secondary outcome) based on changes from baseline in HAM-A score. The MCP-Mod procedure was applied at the 5% significance level, and statistically significant dose-response models were identified separately for each time point.

At week 4, the adjusted critical value was 2.032 and the candidate dose-response models of sigmoid Emax 1 through 4, Emax, and linear observed absolute test statistics of 3.294, 3.149, 3.133, 1.448, 3.471, and 2.971, respectively. Therefore, 5 of the 6 models, including the linear, Emax, and multiple sigmoid Emax variants, met the significance threshold. Among these, the Emax model and 2 sigmoid Emax models were identified as the best fitting based on having the largest test statistic and achieving model convergence.

At week 8, the adjusted critical value was 2.041 and the candidate dose-response models of sigmoid Emax 1 through 4, Emax, and linear observed absolute test statistics of 2.099, 2.922, 2.938, 2.254, 2.690, and 2.778, respectively. Therefore, all 6 models met the significance threshold, with the linear, Emax, and sigmoid Emax models selected as the best fitting. These selected models at each time point were incorporated into an Akaike information criterion–weighted model-averaging procedure to estimate the dose-response curve and calculate the minimally efficacious dose, which was found to be between 56.7 µg and 85.1 µg (eFigure 2 in Supplement 3).

For the primary outcome, the dose-response relationship assessed using the MCP-Mod method for change in HAM-A score at week 4 was statistically significant for the 100-µg and the 200-µg dose groups vs placebo (least-squares mean difference, −5.0 points [95% CI, −9.6 to −0.4 points] with 100 µg of MM120 and −6.0 points [95% CI, −9.8 to −2.0 points] with 200 µg of MM120) but the 25-µg and 50-µg dose groups did not reach significance vs placebo (least-squares mean difference, −1.2 points [95% CI, −6.0 to 3.5 points] with 25 µg of MM120 and −1.8 points [95% CI, −7.6 to 4.0 points] with 50 µg of MM120) (Table 2).

For the key secondary outcome, the dose-response relationship assessed using the MCP-Mod method for change in HAM-A score at week 8 was statistically significant for the 100-µg and the 200-µg dose groups vs placebo (least-squares mean difference, −2.9 points [95% CI, −5.4 to −0.5 points] with 100 µg of MM120 and −5.4 points [95% CI, −9.5 to −1.4 points] with 200 µg of MM120) but the 25-µg and 50-µg dose groups did not reach significance vs placebo (least-squares mean difference, −0.8 points [95% CI, −1.7 to 0.2 points] with 25 µg of MM120 and −1.4 points [95% CI, −3.1 to 0.2 points] with 50 µg of MM120) (Table 2).

The mean changes from baseline for the HAM-A, CGI-S, and MADRS scores appear in eFigure 3 in Supplement 3. To enhance interpretation, the raw per participant changes from baseline in HAM-A score appear in Figure 2. The mean changes from baseline in the HAM-A, CGI-S, and MADRS scores appear in Figure 3.

The secondary outcome analyses demonstrated that the 100-µg and 200-µg dose groups yielded statistically significant reductions in HAM-A score compared with placebo at week 4 (Table 2, Figure 3A, and eFigure 3A in Supplement 3), which was consistent with estimates of the minimally efficacious dose (eFigure 2 in Supplement 3). For the secondary outcome at week 4, the 100-µg and the 200-µg dose groups significantly reduced HAM-A score vs placebo (the least-squares mean difference was −7.6 points [95% CI, −11.8 to −3.4 points; P < .001; Cohen d = 0.88] with 100 µg of MM120 and was −5.5 points [95% CI, −9.7 to −1.3 points; P = .01; Cohen d = 0.64] with 200 µg of MM120) (Table 2, Figure 3A, and eFigure 3A in Supplement 3).

For the secondary outcome at week 8, compared with placebo, the reduction in HAM-A score was shown with a least-squares mean difference of −4.3 points (95% CI, −8.9 to 0.3 points; P = .06; Cohen d = 0.48) for 100 µg of MM120 and the reduction was −5.5 points (95% CI, −10.2 to −0.8 points; P = .02; Cohen d = 0.60) for 200 µg of MM120. At week 12, compared with placebo, the reduction in HAM-A score was shown with a least-squares mean difference of −7.7 points (95% CI, −12.6 to −2.8 points; P = .003; Cohen d = 0.81) with 100 µg of MM120 and the reduction was −7.4 points (95% CI, −12.4 to −2.4 points; P = .004; Cohen d = 0.77) with 200 µg of MM120.

The secondary outcome of HAM-A response (defined by a 50% reduction in total score) at week 12 was achieved by 65.0% of participants in the 100-µg dose group and by 62.5% in the 200-µg dose group compared with 30.8% of participants in the placebo group and the remission rates at week 12 were 47.5%, 45.0%, and 20.5%, respectively (eFigure 4 and eTable 3 in Supplement 3).

There were no significant reductions found in the secondary outcome analyses for the 25-µg and 50-µg dose groups for change in HAM-A score compared with placebo at weeks 4 through 12. At week 4, compared with placebo, the reduction in HAM-A score was shown with a least-squares mean difference of −3.4 points (95% CI, −7.7 to 0.9 points; P = .12) with 25 µg of MM120 and the reduction was −0.9 points (95% CI, −5.2 to 3.3 points; P = .66) with 50 µg of MM120. At week 8, compared with placebo, the reduction in HAM-A score was shown with a least-squares mean difference of −2.9 points (95% CI, −7.6 to 1.8 points; P = .23) with 25 µg of MM120 and the reduction was 0.85 points (95% CI, −3.9 to 5.6 points; P = .72) with 50 µg of MM120. At week 12, compared with placebo, the reduction in HAM-A score was shown with a least-squares mean difference of −4.3 points (95% CI, −9.3 to 0.8 points; P = .10) with 25 µg of MM120 and the reduction was −3.2 points (95% CI, −8.2 to 1.9 points; P = .22) with 50 µg of MM120.

The statistically significant improvements in illness severity (measured using the CGI-S score) were observed on day 2 and at weeks 2 through 12 in the 100-µg and 200-µg dose groups, but not in the 25-µg or 50-µg dose groups or in the placebo group. At week 1, all MM120 dose groups demonstrated significant improvements, but the placebo group did not (Figure 3B and Table 2 and eFigure 3B, eFigure 5, and eTable 4 in Supplement 3).

Significant improvement in depressive symptoms (as measured by MADRS score) was observed from weeks 1 through 12 in the 100-µg and 200-µg dose groups, but not in the 25-µg and 50-µg dose groups (Figure 3C and Table 2 and eFigure 3C and eTable 4 in Supplement 3).

The mean scores at baseline and at week 12 for the HAM-A, the CGI-S, the Patient Global Impression of Severity, and the MADRS appear in eTable 5 in Supplement 3. Additional secondary outcomes, including changes in scores for the Sheehan Disability Scale, the EuroQol 5-dimension 5-level questionnaire, the Pittsburgh Sleep Quality Index, and the Arizona Sexual Experiences Questionnaire appear in eTable 6 in Supplement 3. The analyses of the treatment response by sex and other participant characteristics were not conducted due to insufficient statistical power to evaluate meaningful differences within dosing groups.

Safety

The rate for 1 or more treatment-emergent adverse events (TEAEs) was 76.9% in the 25 µg of MM120 group, was 90.0% in the 50-µg group, was 97.5% in the 100-µg group, and was 100% in the 200-µg group vs 56.4% in the placebo group (Table 3). The rate for any treatment-related TEAE was 74.4% in the 25 µg of MM120 group, was 80.0% in the 50-µg group, was 95.0% in the 100-µg group, and was 100% in the 200-µg group vs 41.0% in the placebo group. Of all the AEs, 98.5% occurred on the dosing day and most resolved by the end of the dosing session.

On the dosing days, the most frequent AEs reported in the 100-µg and 200-µg dose groups were visual perceptual changes (illusion, pseudo-hallucination, hallucination) in 92.5% and 100% of participants, respectively, nausea in 40.0% and 60.0%, headache in 35.0% and 27.5%, and euphoric mood in 27.5% and 15.0%. Additional common (≥10% participants) AEs across all treatment groups appear in eTable 7 in Supplement 3. Most of the AEs were mild to moderate in severity; 2 severe AEs of “feeling intoxicated” were reported on dosing day (1 participant each in the 50-µg and 100-µg dose groups), and both resolved by the end of the dosing session.

One SAE occurred during the study when a participant in the 50-µg dose group was hospitalized for a panic attack that occurred 97 days after the dosing day and was deemed unrelated to the study drug. Five participants (2.5%) withdrew due to AEs (2 participants each in the 25-µg and 50-µg dose groups and 1 participant in the 100-µg group).

No participants engaged in self-injurious or suicidal behavior, which was assessed using the C-SSRS and AE reporting. There were no reports of suicidal ideation with intent or a plan during the sessions or throughout the trial. Participants with a history of suicidal ideation showed no increase in severity.

Based on the C-SSRS assessments, suicidal ideation with any methods (no plan) without intent to act was reported by 2 participants (1 in the 25 µg of MM120 group and 1 in the placebo group), both of whom had a history of similar or more severe suicidal ideation compared with at screening. Additional details on these reports are provided at the end of the eMethods in Supplement 3.

Resolution of acute effects from MM120 was observed in more than 97% of participants at all MM120 doses by 12 hours after the dose was administered (eFigure 6 in Supplement 3). In the 100-µg dose group, 45.0% of participants met criteria to safely end the session at 8 hours, 70.0% at 9 hours, 87.5% at 10 hours, 90.0% at 11 hours, and 97.5% at 12 hours. All participants remained at the study site until 12 hours after the dosing session regardless of meeting criteria to safely end the session per the trial protocol (Supplement 1).

Limitations

Several study limitations should be considered. First, to minimize confounding from psychotherapeutic effects, study-specific training emphasized that dosing session monitors should not engage in psychotherapeutic interventions before, during, or after the session. Two dosing session monitors were present at all times during the dosing session to promote protocol adherence and further prevent intentional or inadvertent psychotherapy. Despite dosing session monitors explicitly not providing a systematic psychotherapeutic intervention, confounding effects of interpersonal interactions remain a possibility. Further details on the training of the dosing session monitors appear in the eMethods in Supplement 3. Although the dosing session monitors were not assessed for a personal history related to psychedelics, it is possible that if such a history were present, it could inform their interactions with participants and further affect the outcomes.

Second, the perceptible drug effects of MM120 may have led to unblinding, which could have influenced the outcomes. All doses of MM120 produced transient perceptual, cognitive, and emotional effects leading to participant unblinding. To mitigate the effects of rater unblinding, independent central raters blinded to the protocol, treatment allocation, and visit number assessed the primary efficacy outcome. Central rater blinding was preserved in more than 80% of HAM-A assessments,³¹ despite approximately 85% of participants correctly identifying assignment to active treatment. In addition, even though more than 65% of participants accurately identified their assignment as placebo recipients, a robust placebo effect was observed that was comparable with or exceeded trials of anxiolytics.³⁸ Despite unblinding occurring at all doses, the 2 lower doses of MM120 did not demonstrate efficacy, suggesting the observed score changes with the higher doses were not the result of participant unblinding. Although participant unblinding is frequently raised as a substantial limitation in trials of psychedelic drugs due to the nature of the perceptual effects,³⁹ it is also observed in other psychiatric drug classes.^40,41 Even though participants with recent or excessive lifetime psychedelic drug use were excluded per the trial protocol, and the enrolled population was well matched to population estimates for lifetime psychedelic use, the total number of enrolled participants with a history of drug experience was not analyzed, presenting the possibility that a participant’s personal history could influence their expectancy, unblinding, or measured efficacy.

Third, the primary outcome and key secondary outcome were strictly controlled for type I error, but the other hypothesis tests were not adjusted for multiple comparisons. Fourth, the dropout rate, though similar across groups, was slightly higher than in studies of similar molecules.^17,18 This higher dropout rate may be due to the 12-week duration of the trial, the required discontinuation of GAD medications, and the absence of an extension phase with opportunities for open-label treatment. Fifth, the participants were physically healthy with limited psychiatric comorbidities, potentially enhancing the tolerability of transient AEs. Sixth, the process of tapering participants from prohibited psychoactive medications ensured MM120 was assessed as a monotherapy; however, we did not address concomitant use of other pharmacotherapies.

Seventh, some participants engaged in psychotherapy outside the trial, which may have affected dosing-day experiences and efficacy outcomes, though this was the case for only 18% of participants. Eighth, the study population was inclusive in terms of sex and age but was racially and ethnically less diverse than the general population of people with GAD. The results may not accurately represent the effects of MM120 in a more diverse population.