A Phase 2 Proof‐of‐Concept, Randomized, Placebo‐Controlled Trial of CX‐8998 in Essential Tremor

Abstract Background Available essential tremor (ET) therapies have limitations. Objectives The objective of this study was to evaluate CX‐8998, a selective T‐type calcium channel modulator, in essential tremor. Methods Patients 18–75 years old with moderate to severe essential tremor were randomized 1:1 to receive CX‐8998 (titrated to 10 mg twice daily) or placebo. The primary end point was change from baseline to day 28 in The Essential Tremor Rating Assessment Scale performance subscale scored by independent blinded video raters. Secondary outcomes included in‐person blinded investigator rating of The Essential Tremor Rating Assessment Scale performance subscale, The Essential Tremor Rating Assessment Scale activities of daily living subscale, and Kinesia ONE accelerometry. Results The video‐rated The Essential Tremor Rating Assessment Scale performance subscale was not different for CX‐8998 (n = 39) versus placebo (n = 44; P = 0.696). CX‐8998 improved investigator‐rated The Essential Tremor Rating Assessment Scale performance subscale (P = 0.017) and The Essential Tremor Rating Assessment Scale activities of daily living (P = 0.049) but not Kinesia ONE (P = 0.421). Adverse events with CX‐8998 included dizziness (21%), headache (8%), euphoric mood (6%), and insomnia (6%). Conclusions The primary efficacy end point was not met; however, CX‐8998 improved some assessments of essential tremor, supporting further clinical investigation. © 2021 The Authors. Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society. This article has been contributed to by US Government employees and their work is in the public domain in the USA.

efficacy, safety, and tolerability of CX-8998 in patients with moderate to severe ET.

Methods
The design of T-CALM (ClinicalTrials.gov: NCT03101241) was previously described 13 and is briefly summarized here. Ethical conduct was consistent with regulatory guidelines.

Study Design and Participants
T-CALM was a phase 2 multicenter double-blind, randomized, placebo-controlled trial of CX-8998, titrated to a target dosage of 10 mg twice daily (20 mg/day), for a total of 28 days, in patients with moderate to severe ET. Patients were 18-75 years of age and diagnosed with classic bilateral ET 14 before age 65. Eligible participants had tremor severity score of ≥2 in ≥1 arm during any of the 3 maneuvers of The Essential Tremor Rating Assessment Scale performance subscale (TETRAS-PS) item 4 (maneuver 1, upper limbs held forward and horizontally; 2, upper limbs extended laterally and horizontally with elbows flexed, and hands positioned close to each other near chin; 3, finger-nose or finger-chin movements) and TETRAS-PS total score ≥ 15 at screening. Use of a stable dosage of a single antitremor medication (with the exception of primidone) was permitted during the study.

Procedures
Patients were randomized 1:1 to receive CX-8998 or placebo, stratified by concomitant use of antitremor medication and study site.
Patients received titrated dosages of CX-8998 during a 4-week double-blind dosing period (4, 8, and 10 mg twice daily during weeks 1, 2, and 3-4, respectively) and were evaluated for safety and efficacy on days 15 (beginning in week 3) and 28 (end of week 4). The dose of study medication could be reduced to the next lower level, if needed; only 1 dose reduction was allowed. Target dose and titration schedule were based on the tolerability profile of CX-8998 (immediaterelease formulation) in a prior clinical study. 15

Outcomes
The Essential Tremor Rating Assessment Scale (TETRAS) comprises a 9-item performance subscale (TETRAS-PS) and a 12-item ADL subscale (TETRAS-ADL). 16 Each patient's TETRAS-PS assessment was scored by investigators in real time at select study visits and by 1 of 3 independent raters of video recordings made during the on-site ratings. Both sets of scores were analyzed using identical methodology.
The primary end point was TETRAS-PS change from baseline to day 28 scored by independent video raters. TETRAS-PS change from baseline to day 28 was also scored in person by investigators. Secondary end points included TETRAS-ADL change from baseline to day 28, rated by patients during an investigator-led interview, and accelerometry with Kinesia ONE. 17,18 Exploratory end points included change from baseline to day 15 and day 28 in TETRAS total score (investigator-rated TETRAS-PS plus TETRAS-ADL); change from baseline to day 15 in TETRAS-PS, TETRAS-ADL, and Kinesia ONE; and ratings of improvement on day 15 and day 28 measured by Clinical Global Impression of Improvement (CGI-I) and Patient Global Impression of Change (PGIC).
Safety assessments included treatment-emergent adverse events (TEAEs), physical examination, neurologic examination, vital signs, clinical laboratory tests, electrocardiogram, and Columbia Suicide Severity Rating Scale.

Statistical Analysis
Approximately 92 patients were planned for enrollment to ensure 86 patients for efficacy analyses. A sample size of 43 patients per treatment group had ≥90% power to detect at least a 5.5-point difference between CX-8998 and placebo for the primary end point, assuming a standard deviation (SD) of 7.5 and α = 0.05. 19 This calculation was based on the Wilcoxon-Mann-Whitney test for 2 independent means and assumed normal distributions for each treatment group with a common, but unconfirmed, SD.
The primary efficacy end point was analyzed with an analysis of covariance model, with fixed effects for treatment, antitremor medication use, study site, and baseline TETRAS-PS score. Secondary and exploratory end points were analyzed similarly.
The intention-to-treat (ITT) analysis set included all randomized subjects. The full analysis set, used for efficacy assessments, included all patients who received any study drug and had both a baseline assessment and ≥1 postbaseline efficacy assessment. The safety analysis set included ITT patients who received any study drug.
About half the population (47%) was female. Mean ± SD age was 63 ± 10.2 years. Mean ± SD time since onset of ET was 23 ± 16.0 years ( Table 1). The treatment groups were matched for most baseline characteristics. Compliance with study drug administration was comparable for CX-8998 (99.3%) and placebo (97.7%).

Efficacy
The difference between treatment groups on change from baseline to day 28 in TETRAS-PS assessments scored by independent video raters (primary end point) was not statistically significant (least-squares [LS] mean ± SE, −1.8 ± 0.81 for CX-8998 vs −2.3 ± 0.78 for placebo; P = 0.696; Fig. 1A). In contrast, TETRAS-PS assessments rated by investigators in person showed improvements with CX-8998 versus placebo on day 28 (LS mean ± SE changes from baseline, −4.8 ± 0.80 for CX-8998 vs −2.8 ± 0.77 for placebo; P = 0.017; Fig. 1B). Intraclass correlation coefficient (ICC) was calculated from a subset of data in which 4 videos were each scored by 4 independent video raters and 1 investigator. The ICC among the video raters was 0.80. When investigator ratings were included in the analysis, the ICC was reduced to 0.60; although this was a limited data analysis, this may be reflective of the discrepancies between video and investigator raters.
CX-8998 demonstrated improvement versus placebo on the CGI-I on day 28 (LS mean ± SE, 1.0 ± 0.13 vs 0.4 ± 0.13; LS mean difference, 0.6; 95% CI, 0.3-0.9; P = 0.001) and PGIC on day 15 (LS mean ± SE,  Baseline was defined as the last nonmissing value that was obtained before or ≤15 minutes after initiation of study drug. c TETRAS total score is sum of TETRAS-PS subscale total score (independent video rated) and TETRAS-ADL subscale score. d Use of primidone at screening was permitted; however, the screening period was to be extended by 2 weeks, for a total of 6 weeks, and primidone was to be discontinued. e Screening until baseline. 0.9 ± 0.17 vs 0.3 ± 0.16; LS mean difference, 0.7; 95% CI, 0.2-1.1; P = 0.003). On day 28, more patients treated with CX-8998 were rated minimally/much/very much improved on the CGI-I and PGIC, compared with placebo (Fig. 1E,F; Table S1). Kinesia ONE scores (triaxial accelerometry and gyroscopy) were similar with CX-8998 and placebo on days 15 and 28 (Table S2).

Discussion
T-CALM was a proof-of-concept study evaluating CX-8998 in patients with moderate to severe ET. The primary efficacy end point, change from baseline to day 28 on TETRAS-PS scored by independent video raters, was not met. However, CX-8998 improved symptoms of ET on a closely related objective measure (investigator-rated TETRAS-PS), as well as TETRAS-ADL, TETRAS total scores, and other patient-reported and clinician-reported end points.
The dual approach to TETRAS-PS scoring aimed to identify the optimal rating methodology for late-stage clinical development. 20 The independent video rater assessment was selected for the primary end point analysis, as it was hypothesized to mitigate investigator bias and variability. The traditional methodology used scores from site investigators who observed patients in real time with the advantage of 3-dimensional angle and depth perception. Compared with the site investigators, the independent video raters consistently scored all but 1 TETRAS-PS subitem lower (data not shown), particularly subitems associated with limitations in videography (face, voice, lower limb, and trunk). 16 In the CX-8998 group but not the placebo group, this effect translated into smaller change scores with video raters versus on-site investigators. These findings may have implications for future clinical trials.
For example, adjustments may improve the videographic process for assessing tremor severity, as size perception is altered (items are perceived as smaller) on images viewed through small 2-dimensional computer screens. 21 In addition, in-person assessments of tremor severity potentially should be performed by raters lacking knowledge of the patient's experience, as investigators may be biased by functional unblinding through other patient observations (eg, AEs, TETRAS-ADL, PGIC).
ET has a major debilitating effect on ADLs. [1][2][3][4][22][23][24][25] Up to 75% of patients with ET experience impairment in ADLs such as eating, drinking, and handwriting. 26 CX-8998 was associated with improvement on TETRAS-ADL, a patient-reported outcome focusing on the functional implications of tremor for ADLs. 16 These results support the use of TETRAS-ADL as a primary end point in future clinical trials.
The Kinesia ONE device, which produces algorithmically derived scores for postural and kinetic tremor in the upper limbs, 17,18 was explored as a potential digital biomarker of tremor severity. However, Kinesia ONE scores were not impacted by CX-8998. This could be related to technical problems (eg, variable finger sensor placement). Future studies with this device and digital biomarkers are essential for further validation of their use in ET clinical trials. 26 The most common TEAEs with CX-8998 were dizziness, headache, euphoric mood, and insomnia and were mild to moderate. Compared with placebo, more patients receiving CX-8998 withdrew because of AEs; no AE leading to discontinuation occurred in >2 patients.
In summary, CX-8998 titrated to 10 mg twice daily ameliorated ET symptoms, although the primary end point was not met. This proof-of-concept study supports further clinical investigation of CX-8998 for the treatment of ET.