Tetrabenazine Versus Deutetrabenazine for Huntington's Disease: Twins or Distant Cousins?

Abstract Background Tetrabenazine is the only US Food and Drug Administration‐approved drug for Huntington's disease, and deutetrabenazine was recently tested against placebo. A switching‐trial from tetrabenazine to deutetrabenazine is underway, but no head‐to‐head, blinded, randomized controlled trial is planned. Using meta‐analytical methodology, the authors compared these molecules. Methods RCTs comparing tetrabenazine or deutetrabenazine with placebo in Huntington's disease were searched. The authors assessed the Cochrane risk‐of‐bias tool, calculated indirect treatment comparisons, and applied the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) framework. Results The evidence network for this report comprised 1 tetrabenazine trial and 1 deutetrabenazine trial, both against placebo. Risk of bias was moderate in both. Participants in the tetrabenazine and deutetrabenazine trials did not differ significantly on motor scores or adverse events. Depression and somnolence scales significantly favored deutetrabenazine. Conclusion There is low‐quality evidence that tetrabenazine and deutetrabenazine do not differ in efficacy or safety. It is important to note that these results are likely to remain the only head‐to‐head comparison between these 2 compounds in Huntington's disease.

blinded, randomized trial is planned. Therefore, we set out to compare TBZ and DEU indirectly using meta-analysis methodology.

Materials and Methods
Our study protocol was registered (PROSPERO CRD42016049199) following the PRISMA-NMA framework. 6 We included randomized controlled trials that compared TBZ or DEU with placebo in patients with HD. The following outcome domains were studied: motor, depression, somnolence, and adverse events (AEs). Severe AEs (SAEs) were classified according to the primary studies, although neither reported a formal definition of an SAE. References were searched in the MEDLINE, Embase, an SAE and CENTRAL databases, the combining (Huntington) with (tetrabenazine OR deutetrabenazine) and applying the Cochrane Highly Sensitive Search Strategy for identifying randomized trials. Studies were evaluated using the Cochrane risk-of-bias tool. Study selection, data collection, and appraisal were done independently in duplicate. Continuous and dichotomous variables were presented as mean differences (MDs) and odds ratios, respectively, both with 95% confidence intervals (95% CIs). Indirect treatment comparison meta-analyses between TBZ and DEU were calculated based on a common comparator using the Bucher method. 7 Confidence in cumulative evidence was assessed using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) Working Group guidelines. 8 These comprise a widely endorsed tool to assess the quality of studies contributing to meta-research that takes into account the domains: risk of bias, inconsistency, indirectness, imprecision, and publication bias, and classifies evidence from high to very low quality, as follows: • High quality: We are very confident that the true effect lies close to that of the estimate of the effect; • Moderate quality: We are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different; • Low quality: Our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect; and • Very low quality: We have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.
A sample-size calculation for a 1:1, parallel equivalence trial was calculated using Stata 14.0 software (Austin, Texas) assuming 80% power, 10% dropout, 0.05 alpha, a standard deviation of 3.5 points on the Unified Huntington's Disease Rating Scale (UHDRS) chorea subscale score, and 20% margin of equivalence of TBZ effect (5 UHDRS chorea score points). 9

Results
In total, 131 references were retrieved, and 2 studies were included. 4,9 Our evidence network, describing how the included studies related to one another, comprised 1 trial that tested TBZ (TETRA-HD; n = 84) and another that tested DEU (FIRST-HD; n = 90), both against placebo (Fig. 1A). The overall risk of bias was moderate in both studies because of attrition and reporting bias (Fig. 1B). In the TBZ arm of TETRA-HD, proportionally more participants withdrew from the study than in the placebo arm; and, in both studies, several important outcome measures, such as quality of life, were missing. In other respects (random sequence generation, allocation concealment, blinding of patient and participants, blinding of outcome assessments, and incomplete outcome data for the DEU trial), the studies were at low risk of bias.
After a detailed review of the methodologies and trial populations, we considered that the included studies were methodologically and clinically similar and comparable on effect modifiers, confirming the transitivity assumption needed to calculate an unbiased, indirect estimate of TBZ versus DEU.

Discussion
Our indirect comparison, as assessed according to the GRADE framework, shows that there is low-quality evidence that TBZ and DEU do not differ in efficacy and safety. DEU appears significantly less prone to depressive symptoms and somnolence, but this observation, which was drawn from indirect analysis of a restricted evidence network, requires validation in a direct, suitably designed trial.
Our analysis must be interpreted with caution overall, because indirect comparisons only provide observational evidence: the power of hypothesis testing relies on between-study heterogeneity, which thankfully was minimal in this case. Furthermore, the power of our computation is limited by the evidence network sample size. 10 If DEU receives licensing authorization, then long-term, phase 4 studies and real-world practice will provide further information on the clinical utility of DEU. Nonetheless, our analysis raises concerns for informed clinical decision making in HD: no clinical trial has recruited over 600 participants; and, to our knowledge, only 1 ongoing trial seeks to compare TBZ and DEU directly: ARC-HD, whose nonrandomized, openlabel, switching design carries a risk of selection, detection, and performance bias. Therefore, the present study seems likely to remain the only feasible and realistic, blinded, head-to-head comparison between TBZ and DEU in HD.