The specific aims of this project are to determine if CBD augmentation compared to placebo: significantly modulates psychotic symptoms, cognitive symptoms, pro-inflammatory, and neurohormone biomarker levels, appetite and eating behavior, metabolic measures, and lastly changes neurophysiological abnormalities.

It is hypothesized that CBD will reduce symptoms, improve neurocognition, reduce markers of inflammation, reduce harmful eating behaviors, and improve information processing. Importantly, predictors, moderators, and mediators of the CBD effects will be examined. A better understanding of the mechanism by which CBD acts in early psychosis and the identification of which individuals are more likely to respond to CBD can inform treatment development and personalize treatment.

The study will be conducted in the UCSD CARE (University of California, San Diego, Cognitive Assessment and Risk Evaluation) Early Psychosis Program where we have a multidisciplinary team of psychiatrists and psychologists with expertise in early psychosis, integrative medicine techniques, biomarkers, and clinical trials. UCSD CARE early psychosis and treatment program is in the Psychiatry Department of the University of California San Diego, California. The CARE early psychosis clinic sees up to 4 new patients per week, generally consisting of a relatively high percentage of patients who would be eligible for this study. The catchment area is the entirety of the multicultural and diverse county of San Diego. The coordinating team is part of UCSD CARE and is comprised of the PI, a clinical psychologist, a graduate student, a clinical research coordinator, and three research associates. The research associates are responsible for identifying potential recruits via outreach and referral screenings, consenting participants, and implementing the test battery. The clinical psychologist and graduate student are responsible for administering the clinical measures. The testing procedures are overseen by the clinical research coordinator who manages the study database and coordinates the study with the outside collaborators. The collaborators on this project are the two funding agencies. The first agency is the Krupp Center of Integrative Research at UCSD co-directed by the statistician consulting on this project. The second agency is the Center for Medicinal Cannabis Research (CMCR) which acts as the IND sponsor, monitors the study progress, provides ongoing assistance to the PI on study design and any issues that arise, and will provide assistance in analyzing the study results. The CMCR will also host the collected data and provide long-term storage for the biological samples collected during the study. The samples will not undergo genetic analysis and will be used only for metabolic and inflammatory analysis and will not be specifically stored for future use.

As outlined in the study design (Fig. 1) and SPIRIT (Fig. 2), the trial will include 120 patients in the early stages of psychosis who will be randomly assigned in a 1:1 ratio to receive 1000 mg of cannabidiol or matching placebo as an add-on to current treatment to in a 8-week double-blind superiority randomized controlled trial to be performed over 3 years. The unblinded study statistician has created a list of participant identification numbers with randomized treatment assignments using the SPSS software prior to the start of the study subject recruitment. Blocked randomization will be used with a random block size (4 or 6). Subjects will be randomized to the study treatment group in a 1:1 ratio. No stratification method will be used. A research associate on the team who is not part of the study will be provided with that subject’s randomization code. Any violations of the inclusion and exclusion criteria, reasons for randomization failure including major protocol deviation, lost to follow-up, voluntary withdrawal, and study termination, will be communicated to the study statistician. This research associate is also the only person who is not blinded and who has access to the list that identifies the unique numbers on the relabeled bottles of the two products. They are then responsible for assigning the bottles to the patient. All clinical staff and participants involved in the study will be blinded with the sole exception of the independent statistician. The design is sufficiently powered to answer questions about the efficacy of CBD as an augmentation to antipsychotics in EP, predictors, mediators, and moderators of treatment response and further our understanding of the mechanism of action of CBD in early psychosis. Symptoms, eating behaviors, metabolic parameters, and hormonal and neuroimmune marker levels will be evaluated at baseline, 3 weeks, and 8 weeks, and neurocognitive and neurophysiologic assessment will be performed at baseline and 7 weeks. We will work closely with the UCSD Center for Medicinal Cannabis which has performed multiple CBD trials. UCSD IRB approval, FDA IND approval, and ClinicalTrials.gov registration are in place. Data will be deidentified and stored in RedCAP, a secure dataset available through UCSD. A data monitoring committee is made up of members of both funding agencies, which has assisted in the development of the database and ensuring compatibility with data forms. There is a Data and Safety Monitoring Board (DSMB) that reviews the progress of the study, recruitment, dropouts, and adverse events of trial participants on an annual basis, the board is made up of an independent group of investigators. Any protocol modifications will be reported to IRB, trial registry, and DSMB.Fig. 1

Study design

In order to be eligible to participate in this study, an individual must meet all of the following criteria:

Between the ages of 16 and 30.

An individual who meets any of the following criteria will be excluded from participation in this study:

Concomitant medical or neurological illness

If clinically significant, untreated symptoms are found to be present at baseline, or are found to develop during the course of the study, participants will be advised to consult with their mental health care provider and will also be provided with referrals to UCSD (e.g., UCSD CARE Early Psychosis Specialty Clinic-clinical arm of the CARE) and appropriate clinicians in the community. These participants will continue to be followed in the study, with information on treatment recorded in the database.

On the initial screening visit, the purpose of the study will be explained by a research associate. Individuals who are interested in participating in the study will receive more detailed information, both orally and with a written informed consent. Potential study participants will be asked to sign a separate release of information as well as the UCSD HIPAA form in case information needs to be obtained or released to a third party. Consenting will be completed by experienced research staff. In case of minor participants, both assent from the adolescent and consent from the parent/guardian will be completed.

Consent forms will be written in language that is comprehensible to individuals who have at minimum an eighth grade education. The consent form will contain information about the nature of the interviews, cognitive testing, fluid sampling, and neurophysiological procedures. Procedures for maintaining the confidentiality of data will also be described.

To assess the comprehension of the information provided, after reading the consent forms, subjects or parent/guardian will be asked to provide a statement of their understanding of the research, the research procedure, any risks or discomfort involved, possible benefits of the study, and their right to withdraw at any time. If a subject indicates an inability to comprehend the provided material or is judged to be unable to comprehend this material, that individual will not be permitted to participate in the research. Subjects will receive copies of the signed consent documents, HIPAA authorization forms, and “The Experimental Subject’s Bill of Rights.”

For participants who start as minors and become of age (18 years old) during the length of the study, efforts will be made to obtain consents from participants who completed the active phase of the study but whose biospecimen continues to be stored. A waiver of consent is requested with the goal to continue storing previously collected and stored biospecimen in the case that participants who have come of age cannot be reached. The continued storage of the biospecimen is considered a minimal risk to subjects, given that strict confidentiality as well as safety of storage procedures will be maintained as described in extensive documentation approved by our IRB. The rights and the welfare of the participant are not adversely impacted given that study procedures will have been completed at that time point, and none of the results of the research using the biospecimen would affect clinical decisions about the individual’s care because the results are analyzed after the fact. The biospecimen storage and additional analysis of biospecimen cannot be carried out without the waiver. There is no plan for genetic and/or molecular analysis of biomarkers, only routine biomarker quantification, without long-term storage for specific future studies. If a waiver is not obtained, biospecimen will need to be destroyed and valuable opportunity for scientific knowledge that can be derived from it will be lost.

The Epidiolex and matching placebo are manufactured and supplied by Jazz Pharmaceuticals. The Epidiolex formulation is a 100-mg/mL solution. The CBD compound will be dosed at 1000 mg/day administered in two divided doses after an initial 1-week lead-in of 500 mg. The dose of CBD was selected based on previous controlled trials that demonstrate the efficacy of CBD in patients with schizophrenia [16]. The active pharmaceutical ingredient in cannabidiol oral solution is a pharmaceutical-grade synthetic cannabidiol manufactured according to the current good manufacturing practice (cGMP). It is an off-white to pale yellow resin or crystal substance that is soluble in several organic solvents with an acid dissociation constant (pKa) of 9.64. The solution is a clear, colorless to pale yellow to brown colored solution filled into a 30-mL amber glass bottle of 100 mg/mL strength (i.e., 3000 mg per container) that will be stored at a controlled room temperature (20 to 25 °C, 68 to 77 °F) at the study center where the investigational product will be dispensed. DEA regulations are followed and detail specific security requirements for storage of the investigational product in a “securely locked, substantially constructed cabinet.” Investigators must notify the DEA of the theft or significant loss of any controlled substances within 1 business day of discovering such loss or theft. Any loss in investigational products must be reported to the sponsor. The matching placebo is identical in texture, color, taste, and packaging and will be administered in the same volume as the CBD, so that parents, participants, and investigators should not be able to detect differences between the two treatments. Upon reception of the Epidiolex and placebo bottles, they are all relabeled with an identical label and each assigned a unique number by a team member who is not part of the study and who will be the only person with access to the list identifying the numbers with the vial contents.

Adherence will be assessed during a weekly check-in, bottle inspection, and CBD blood levels at the beginning, 3 weeks, and end of the trial. Participants will be queried as to whether they have missed doses of antipsychotic or study medication. Ease of study recruitment and attrition from the study will also be documented.

Confidentiality will be maintained by assigning each subject a study number and coding all data collected with that number. Identifying information will not be stored on computer databases and will not be stored with the study subject number. All computer databases are password-protected, and hard copies of all data and records will be stored in locked filing cabinets. All study personnel will be certified to conduct research with human subjects and will be aware of the importance of maintaining strict confidentiality.

Power analyses were conducted allowing for a 20% dropout rate, and a sample size of 120 early psychosis patients (60 per group), measured at 3 time points, was calculated as sufficient to detect a treatment difference in PANSS/CGI total score, biomarker assessments, food intake and metabolic indices, and neurophysiological measures between CBD and placebo on the change from baseline to end of treatment, with a two-tailed 1% significance level and 80% power. We selected a sample size that provided a minimum of 80% power for medium effect sizes based on the anti-inflammatory outcome measures reported in multiple studies [106–108] that show medium effect sizes (Cohen d = 0.47; Hedges g =  − 0.39 in EP) for clinical symptoms.

Ensuring adequate recruitment requires a number of strategies, including providing announcements to local primary care practitioners, including community pediatricians, family practice physicians, educators, clinical social workers, psychologists, and psychiatrists. Ads are also posted on various social media platforms. Additionally, ResearchMatch.org and clinicaltrials.org will be utilized as a recruitment tool for this protocol. We estimate that we will be able to recruit five new patients every month based on past referrals.

Data will be analyzed using a mixed model [109] for all clinical, neurocognitive, biomarker, eating behavior, and electrophysiological markers. This method allows the inclusion of subjects with missing data or those who were terminated early in the study, without relying on data imputation procedures. This method provides an estimate of the individual variability around the population trend, the variability of the individual intercepts (baseline values) and slopes (changes across time), and the correlation between them. The model will include a random intercept, a random effect for assessment time, and fixed effects for comparison groups and group-by-time interaction. A fully saturated treatment-by-time model will be utilized for inference. Co-variance structure will be chosen based on Akaikes Information Criterion (AIC). Random group-level treatment effects will also be evaluated for importance based on the model AIC. Any group-level effects to be incorporated into the model. Denominator degree’s of freedom will be calculated using the Kenward-Roger small sample correction. Data will be analyzed from all randomized subjects on whom we have a baseline assessment and at least one post-baseline evaluation. Hypothesis 5 will be testing using the multiple regression procedure. All tests will be done using the SPSS version 24.

We will explore predictors (moderators) of treatment response including baseline symptom severity, neurocognition, biomarker (cortisol and inflammatory markers), and neurophysiological markers by building hierarchical linear models (HLM). The independent variables are the treatment group, moderator, and the treatment × moderator interaction. Mediation analyses will use a similar linear mixed models approach. Two conditions must be met: (1) correlation between the mediator and treatment and (2) relationship between the mediator and outcome. First, we will test the effect of the treatment group and the group × time interaction on the mediator, and we expect a statistically significant group × time interaction. Second, we will test the effects of mediator by group interaction on the outcome in the model that includes group and time; we expect a statistically significant mediator × time or mediator × group × time interaction.

Number, time, reasons, and grade (severity) of adverse events will be summarized overall and by treatment using mean (SD) or median (IQR), range, and N (%) as appropriate. The proportion of subjects with adverse events will be compared between treatments using the McNemar test. The number of adverse events will be compared between treatments using the Wilcoxon signed-rank test. Serious adverse events are reported to the IRB within 24 h of the occurrence of the event and follow the proper procedures regarding the reporting of expectedness, seriousness, severity, and causality. The study subjects are instructed to report any new unusual symptoms, suspected side effect, or health-related events as soon as they occur during the 7-week trial period. On the weeks where they do not have an in-person visit, the subjects are contacted by phone by the clinical psychologist and asked a series of questions to assess any side effects that they might be experiencing. All serious adverse events are also reported in the annual DSMB report.

Baseline demographic measures and relevant diagnostic characteristics will be summarized as described under general approach. If dropout occurs, the characteristics of participants who withdraw from the study prematurely will be compared to those who complete the study as described under general approach. We used the SPIRIT reporting guidelines for this manuscript [110].