Cannabidiol (Epidyolex®) for severe behavioral manifestations in patients with tuberous sclerosis complex, mucopolysaccharidosis type III and fragile X syndrome: protocol for a series of randomized, placebo-controlled N-of-1 trials

Rationale for the N-of-1 design

Trials in RGNDs pose specific methodological challenges due to comorbidities and rarity of conditions [32, 33]. Additionally, patients with rare disorders require individualized treatments and outcome measures due to their heterogeneity and vulnerability. It is therefore difficult to conduct traditional randomized controlled trials (RCTs) to determine effectiveness. The N-of-1 methodology is an alternative type of RCT, providing rigorous, and the highest level of evidence of treatment effectiveness at an individual level and is consistent with the movement towards personalized care [34, 35]. N-of-1 studies are randomized, controlled, multiple cross-over trials within individual patients [36, 37] and enhance treatment precision when intervention effects are heterogeneous between individuals [38, 39]. In this way, structured and evidence-based treatment decisions can be made for an individual patient. N-of-1 trials are ideal for studying treatment effects that vary among patients [40, 41]. N-of-1 trials utilize repeated measurements within individuals, considering variability among and within patients [42]. Using statistical methods like mixed-effects models accommodates individual variations through specific random effects for each patient. Analyzing repeated measures separately for each patient allows these models to address inter-individual differences, offering a robust way to assess treatment effects in heterogeneous populations. Aggregating the results of several N-of-1 trials in different rare, complex, and heterogeneous disorders yields treatment effect estimates [43, 44] and contributes to the generalizability to future patients with these RGNDs, but potentially also to patients with other RGNDs [45].

Study design and duration

We have used the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) extension for N-of-1 trials (SPENT) checklist that is aligned with the CONSORT (consolidated reporting items for trials) extension for N-of-1 trials (CENT) for developing this N-of-1 protocol [39].

The study will consist of a series of N-of-1 trials followed by an optional open-label extension phase. Each N-of-1 trial is block-randomized, placebo-controlled, and double-blinded with multiple crossovers in a single patient. The trial will start with a baseline period of 2 weeks without any intervention. A variable dose titration phase will follow with a taper period (2 weeks) and washout period (1 week) before starting the trial. Each N-of-1 trial consists of two cycles, each consisting of one period of CBD treatment (A; 6 weeks), one period of placebo treatment (B; 6 weeks), run-in periods (3 weeks), taper periods (2 weeks), and washouts following A and B (1 week) (Fig. 1). The half-life of cannabidiol in plasma is 56–61 h after twice-daily dosing for 7 days in healthy volunteers (Summary of Product Characteristics Epidyolex, https://www.ema.europa.eu/en/documents/product-information/epidyolex-epar-product-information_en.pdf). Taking this pharmacokinetic profile into account, and considering the presence of a buildup phase before the commencement of each cycle, it is expected that a 1-week washout period between cycles will adequately allow for the dissipation of any residual effects, facilitating a clear differentiation between treatment cycles.

The total duration of each trial will last around 1 year. The optional open-label extension phase will be a further 12 months.

Study setting

Prior to the start of the trial, all clinical measures and questionnaires will be completed in the Amsterdam UMC, location AMC. The remaining questionnaires will be completed and study medication will be taken at the participants’ natural setting.

Recruitment

Participants will be recruited through the outpatient clinics for TSC and FXS at ‘s Heeren Loo, the UMC Utrecht and the Erasmus MC, and through patient organizations and the MPS III expert center at the Amsterdam UMC.

Study population

The study population consists of children and adults with TSC and FXS, and children with MPS III, all suffering from severe behavioral manifestations. We aim to conduct a patient-centered trial, allowing for a natural setting and flexibility, including continuation of concurrent therapies.

Inclusion criteria include:

• Minimum age of 6 years old.
• Clinically and/or genetically definite diagnosis of TSC, MPS III or FXS (modified Gomez Criteria, clinical criteria, positive genetic test, or enzyme deficiency).
• Suffering from severe behavioral manifestation with a minimum score of 4 on the Clinical Global Impression – severity (CGI-S) scale [48].
• Stable dose of all psychopharmacological medications or interventions for one month prior to screening and willingness of the participant and legal representatives to maintain the current medication regimen throughout the trial.
• Presence of a consistently available patient caregiver for proxy reports.

Exclusion criteria include:

• Any known or suspected hypersensitivity to cannabinoids or any of the excipients of the Investigational Medicinal Product (IMP), such as sesame oil.
• History of recreational or medicinal cannabis, or cannabinoid-based medications, with three months prior to screening and the patient is unwilling to abstain for the duration of the study.
• History or current evidence of significantly impaired liver function, defined as 1) Alanine aminotransferase (ALT) and/or aspartate aminotransferase (AST) > 5 x upper limit of normal (ULN); 2) ALT or AST > 3 x ULN with concomitant total bilirubin > 2.0 x ULN; or 3) ALT or AST ≥ 3 x ULN with the appearance of fatigue, nausea, vomiting, right upper quadrant pain or tenderness, fever, rash, and/or eosinophilia.
• Pregnancy or breastfeeding. Females of childbearing potential must be willing to use an effective method of contraception from the time of consent until 6 weeks after treatment discontinuation and inform the trial if pregnancy occurs.
• Glaucoma.
• History of general anesthesia in the 4 weeks prior to enrolment.
• Use of any interfering medication within 30 days prior to enrolment or planning to take interfering medication during the trial.
• Any planned major surgery within the duration of the trial.
• Expected inability to undergo blood sampling due to anxiety or resistance.
• Unwillingness or inability to swallow the study drug (or placebo).

In addition to the exclusion criteria, use of valproate should be stable 3 months prior to enrolment.

Sample size

The sample size calculation was based on a summary measures analysis of the treatment effect [49]. The difference between the mean irritability ratings of the Aberrant Behavior Checklist (ABC) in CBD periods and placebo periods was used as a summary measure for the treatment effect in an individual participant. Heussler et al. [9] reported a standard deviation (SD) of 12 points for single ratings on the subscale. This estimate for the SD includes both within- and between-subject variance. The intraclass correlation coefficient (ICC) was assumed to be 0.75. In addition to the estimate for the within-subject variation of the outcome measure, an a priori estimate was needed for the between-subject variation of the treatment effect. Assuming an SD of 3 points for the SD of the random treatment effect, 95% of the subject-specific treatment effects roughly falls within a range of 12 points. Based on the estimate assuming two cycles with six ABC ratings within each period and assuming a two-sided significance level of 5%, a total of 6 participants will yield 80% power to detect a mean difference of 6 points between intervention and placebo periods, corresponding to a mean difference of 0.5 times the SD as reported by Heussler et al. [9]. As effects for pediatric and adult population may differ, separate power analyses were performed for the pediatric and adult cohorts. Per cohort, this amounts to 6 pediatric patients and/or 6 adult patients, with a total inclusion of 12 patients with TSC, 12 with FXS and 6 pediatric patients with MPS III. MPS III is a type of childhood dementia and most patients never reach adulthood [26]. Adult patients experience progressive dementia. This stage of the disease is not associated with severe behavioral manifestations, and therefore we will not include adult patients with MPS III.

Blinding, treatment allocation, randomization

Participants, parents, caregivers, physicians and researchers will all be blinded during the trial. The random allocation sequence will be generated for block randomization in a 1:1 ratio and implemented by the hospital pharmacist and sequentially numbered packages. Unblinding will occur when a participant has completed the two cycles or in case of a serious adverse event (SAE) that cannot be treated without knowing which treatment the patient was receiving. Investigators involved in data analysis will remain blinded until the end of the follow-up period.

Interventions and dosing schedule

Patients will receive a pharmaceutical formulation of highly purified CBD derived from Cannabis sativa L. (100 mg/mL) oral solution (Epidyolex® [Jazz Pharmaceuticals]) alternately with a placebo distributed by the Amsterdam UMC hospital pharmacist. CBD reduced TSC-associated seizures versus placebo with similar efficacy between the 25 and 50 mg/kg/d doses [20]. Given that the safety profile of 25 mg/kg/d was superior to 50 mg/kg/d, the lower dose range suggests a superior benefit-to-risk ratio. Standard rules for the use of CBD are in force. Participants can continue their psychopharmacological medications.

Outcome measures

The primary outcome is a change on the irritability subscale score of the Aberrant Behavior Checklist (ABC-I) during active interventional periods compared to placebo periods.

Secondary outcome measures include:

• Total ABC [50];
• CGI [48];
• Syndrome-specific outcome measures, including the TSC-specific patient-reported outcome measure (TSC-PROM) [51] and the Sanfilippo Behavior Rating Scale (SBRS) [52];
• Pediatric Quality of Life Inventory (PedsQL) [53];
• Anxiety, Depression, and Mood Scale (ADAMS) [54];
• Social Communication Questionnaire (SCQ) [55];
• Social Responsiveness Scale (SRS)-2 (when appropriate) [56];
• Short Sensory Profile (SSP-NL) [57];
• Parenting Stress Questionnaire (OBVL) [58];
• Goal Attainment Scaling (GAS) [59];
• Personal Questionnaire (PQ) [60];
• Focal and generalized seizure frequency;
• Adverse effects;
• Hepatic enzyme levels.

Baseline clinical characteristics, demographic information, medical history, results on CYP enzymes, and information regarding diagnosis will be collected and recorded in detail to enable better interpretation of results and explore factors associated with treatment response, because of the heterogeneity of the population and diverse neurobiological and behavioral phenotypes [47]. Collecting data on CYP enzymes has already been best practice and recommended to measure for refractory or difficult to treat behavioral manifestations in this population. The shortened version Vineland Adaptive Behavior Scales-III (VABS-III) [61] will be filled out by the clinicians to determine whether the SRS-2 could be filled out during the trial [62]. Optionally, functional excitation-inhibition electroencephalogram (fE/I EEG) recordings will be performed at baseline to detect changes in brain resting-state EEG for stratification of responsiveness post hoc. The fE/I EEG is optional considering the potential burden of this additional assessments, particularly in the context of the n-of-1 design. This study utilize specialized software tools, such as Matlab and the Neurophysiological Biomarker Toolbox, to preprocess and analyze fE/I EEG data. This comprehensive analysis allows us to evaluate various resting state parameters, such as power spectra, coherence, and E/I ratios [63]. fE/I EEG assessments are integral to our research, providing valuable insights into the functional connectivity and balance of neural networks, particularly relevant in individuals with neurodevelopmental disorders. Our investigation focuses on understanding the neural mechanisms underlying these conditions and their potential responses to therapeutic interventions, with a particular emphasis on E/I ratios and other EEG parameters.

Trial procedures

During the N-of-1 trial, the final CBD dosage as determined by the dose titration phase or placebo will be administered twice daily. During washout periods, no study medication will be taken.

Safety evaluation

The authors would like to thank the Dutch patient advocacy organizations (Stichting Tubereuze Sclerosis Nederland, Fragiele X Vereniging Nederland) for their feedback on the study protocol. The Amsterdam UMC and Erasmus MC are member of European Reference Network (ERN) ITHACA. The Amsterdam UMC is member of MetabERN. The UMC Utrecht is member of ERN EpiCare.