The protocol for the Cannabidiol in children with refractory epileptic encephalopathy (CARE-E) study: a phase 1 dosage escalation study

As there is no available pediatric pharmacokinetic data for the cannabinoids including CBD and THC, the dosage regimen used in this study is extrapolated from CBD dosages previously described in the literature [2‐6]. Consideration is made of the fact that the study product is derived from a whole plant extract that contains Δ⁹-THC among other potentially biologically active cannabinoids and terpines.

In Jacobson and Porter’s report, most children who had a positive response to CBD were taking a dose ranging from 8 to 14 mg/kg/day [2]. Devinsky and Thiele used a dose of 20 mg/kg/day in their participants randomized to receive study drug but this was a purified CBD product with negligible concentrations of Δ⁹-THC [5, 6]. Tzadok’s study participants received a CBD dose of either < 10 mg/kg/day or 10–20 mg/kg/day provided in the form a CBD-enriched Cannabis extract [4].

A secondary study objective is to determine the relationship between dose escalation and steady state trough concentrations of bioactive cannabinoids, and if possible, relate these levels with therapeutic and adverse effects. To achieve this objective a liquid chromatography-mass spectrometry (LC-MS/MS) method was validated in accordance with the United States FDA guidelines [12, 13]. Blood collected into lithium heparin Barricor vacutainers ® (BD Canada, Mississauga, ON) at each visit will be centrifuged (10 min at 1500 rpm), the plasma aliquoted into clearly labeled microcentrifuge tubes, and placed at − 80 °C until analysis. Plasma concentrations of THC, CBD and THC-OH (11-hydroxy-THC) in participant plasma samples will be determined by LC-MS/MS analysis. Briefly, stock solutions (1 mg mL^− 1) of cannabinoids and their respective stable isotope labeled internal standards (Cerilliant Corp., Round Rock, TX) will be prepared in methanol and stored at − 20 °C. Working solutions will be prepared by serial dilution of the stock solution in blank human plasma to produce appropriate standard calibration curves. Acceptance criteria for each analytical run will be based on low, medium, and high concentration quality control (QC) standards. Calibration and QC samples will be prepared on each day of sample analysis. A linear least-squares regression analysis using 1/X² as weighting factor will be conducted to determine the linearity of the calibration curve. Plasma sample extraction involves the addition of 10 μL of the internal standard working solution and 600 μL of cold acetonitrile to 200 μL plasma, followed by vortex-mixing and centrifugation at 20,000 g for 10 min at 4 °C. 700 μL of supernatant is dried under filtered air for 15 min at 37 °C. Samples are reconstituted using 200 μL mobile phase. Supernatant will be transferred to HPLC inserts and 5 μL injected onto a Zorbax Eclipse XDB-C18 narrow bore 2.1 × 12.5 mm 5 μm guard column and Zorbax Eclipse XDB-C8 narrow bore 2.1 × 12.5 mm 5 μm guard column with column temperature maintained at 30 °C. The cannabinoids are separated using an Agilent series 1290 binary pump (Agilent Technologies, Mississauga, ON, Canada) with an online degasser and auto sampler set at 4° and a mobile phase of 80% methanol and 20% Solution B (0.1 mM ammonium formate) at a flow rate of 250 μL/min. Injections will occur at 13.5 min intervals and will include linear gradients to 90% methanol 10% Solution B at 3.5 min to 10 min and return to 80% methanol: 20% Solution B from 10 min to 10.5 min.

The cannabinoids will be detected with an ABSciex 6500 QTRAP mass spectrometer (AB Sciex, Concord, ON, Canada) in positive ion mode. Multiple reaction monitoring (MRM) will be used to quantify the cannabinoids and the peak areas will be summed through use of MultiQuant 3.0.1 Software. The ratio of peak areas of the cannabinoids to their respective internal standards will be plotted against the nominal concentrations to construct the calibration curve and the concentrations of the cannabinoids determined by interpolation.

At each visit participants will have laboratory assessment of blood components to evaluate hepatic, renal, or hematopoietic toxicity performed at their local hospital laboratory. The tests performed include: a complete blood cell count panel with automated three or five part cell differential, electrolytes, glucose, creatinine, urea, alanine transaminase, aspartate transaminase, albumin, gamma glutamyl transferase and lipase. Adverse events from each participant will be assessed as laboratory results that exceed the local laboratory age-specific reference intervals. If participants are on a ketogenic diet during the study, then urine ketone testing will be performed to assess the consistency of the ketosis at each visit.

Participants will remain on pre-existing anticonvulsant medications throughout the cannabis oil study period. Serum specimens will be collected from participants at each visit and trough levels of serum anticonvulsant medications will be determined by LC-MS/MS by the Roy Romano Provincial Laboratory Regina, SK, Canada. Serum specimens were collected and stored at − 20 °C prior to analysis. Adverse events will be counted if participants require a change in anticonvulsant medication during the trial either to maintain trough levels in the therapeutic range.

The instrument we have chosen is the Quality of Life in Childhood Epilepsy (QOLCE-55) [14]. The QOLCE is a parent/proxy-completed measure of health-related quality of life specifically developed for children with epilepsy. It has several subsections containing multiple items, as well as a series of global ratings. The original tool was designed for individuals between 4 and 18 years of age which is one of the broadest age ranges for a tool of this kind. The tool allows for the rater to indicate that an item is not applicable if its content is above the age or developmental level of the child being rated. This makes the QOLCE potentially robust in the face of issues such as lower age and intellectual disability.

The QOLCE-55 shows good internal consistency and criterion-related validity as well as adequate to good test-retest reliability, depending on the subtest or item involved [14‐16]. Areas covered include physical features (including physical limitations and fatigue), well-being (including depression, anxiety, helplessness and self-esteem), cognition (including attention, memory, language and general cognition), social engagement (including interactions, activities and stigma), and behavior. The QOLCE has also been shown to be sensitive to seizure activity and other clinical and psychosocial variables associated with epilepsy [14] and to benefits from treatments such as surgery [17]. Finally, the QOLCE has been used in the study of epileptic conditions with associated cognitive delays and Intellectual Disability and has already shown its utility in samples with Intellectual Disabilities [18]. While the QOLCE-55 was not exclusively positive in the wording of its items, most items were positively stated, making for less distress on the side of those completing the measure [19].

Ratings on the QOLCE are made on a 5-point scale with 1 titled “very often” and 5 titled “never.” Reversed items are recoded when scoring such that higher scores mean more positive outcomes. These scores are then recoded as follows: 1 = 0, 2 = 25, 3 = 50, 4 = 75, and 5 = 100. The mean for each of the subscales is then found by adding these values together and dividing by the number of items not marked Not Appropriate. The total score for the scale is the unweighted mean of the four subscales.

As well, for the purposes of our study we added 13 additional items based on reports from parents. Additional items covered sleep (including being drowsy), verbal and nonverbal communication, use of books, awareness of surroundings, interpersonal interactions with children and adults, and irritability. These additional items are scored as other QOLCE items and are summed into their own total score as well as being looked at individually.

Seizure monitoring will be used to determine how treatment with the study compound affects seizure frequency duration. Caregivers will be asked to track the frequency and duration of their child’s three most frequent types of seizures on a daily basis using a study diary. In order for the study to remain consistent, the caregivers will track the same three types of seizures throughout the study. Seizures that occur in a cluster will be counted as one seizure although the duration of the cluster and number of seizures per cluster will be recorded. Although dialeptic seizures are not included as part of the inclusion criteria for the study, caregivers will be encouraged to record the frequency of dialeptic seizures if their child experienced them frequently.

Caregivers will be asked to track their child’s use of rescue medication. This will determine whether treatment with the study compound has any influence on use of rescue medication. Caregivers will record the medication used, the dosage used, and the number of times it was administered.

All data will be descriptively analyzed using means, standard deviations, frequencies (where appropriate), and 95% confidence intervals. The sample size of 28 participants is sufficient for an initial phase 1 safety and tolerability study, but is too small for precise estimation of steady state levels of biologically active cannabinoids at each dose and for definitive assessments of efficacy. Trends will be examined and a medical statistician will assist with statistical and trend analysis of the data. Complete, specific details of the statistical analysis will be described and fully documented in the Statistical Analysis Plan (SAP) after completion of data collection.