Treatment of tardive dyskinesia with tetrabenazine or valbenazine: a systematic review

The PubMed and Embase databases were searched for English language publications in the date range of 1 January 1980–31 March 2017. Since reports have already been published that have reviewed earlier studies of TD with tetrabenazine [27], studies published prior to 1980 were excluded from analysis with the intention of minimizing historical bias and enhancing comparability in methodology with the more recent valbenazine trials. Search strings for each database were ‘valbenazine AND tardive dyskinesia’ and ‘tetrabenazine AND tardive dyskinesia’. Duplicate results from these four separate searches were removed.

Studies were selected for inclusion if they were a randomized controlled trial, single-arm study, cohort study, case series (with more than or equal to ten TD cases) or retrospective chart review. Case reports with less than ten TD cases, meta-analyses, reviews, animal studies, modeling studies, pharmacokinetic studies, chromatography studies, child/pediatric studies, guidelines, articles and letters were excluded. Results from the literature searches (i.e., study titles and/or abstracts) were screened to remove irrelevant studies based on the exclusion criteria. Results of this screening process were then further assessed by evaluating full-text articles for eligibility.

Full-text versions of selected studies were assessed to determine study design, sample size, study sites (single, multiple), comparator(s), efficacy and safety outcomes. For studies that reported the treatment of several movement disorders, only efficacy results pertaining to TD were reviewed and summarized. Based on available information, safety results were reviewed and summarized for the overall study population (i.e., any movement disorder) or only for the TD population. A meta-analysis of treatment effects was originally planned. However, as reported below, differences in study design did not allow for any meaningful comparisons across studies. Therefore, only a descriptive review of the studies is provided.

The initial database searches yielded 487 results, 83 of which were duplicates and thus excluded from screening. Screening of titles and abstracts resulted in the exclusion of an additional 391 studies. A total of 13 studies met criteria for this systematic review upon screening and full-text articles were assessed for eligibility (Figure 1). Four studies were identified for valbenazine: one open-label extension to a Phase II study [28]; two randomized, double-blind, placebo-controlled trials [26,29]; and one double-blind extension to a Phase III study [30]. Eligibility assessment of these studies resulted in the exclusion of the open-label extension study [28]. Because this open-label study was only presented orally in a platform session at the 2016 American Academy of Neurology (AAN) Annual Meeting, information about the methods and results were limited. Nine studies were identified for tetrabenazine: one double-blind, crossover study [31]; and eight open-label studies, which included two case series [32,33], three long-term studies [34–36] and three retrospective chart reviews [37–39]. Full-text review resulted in the exclusion of one open-label case series [32] because only four of the 14 TD patients were treated with tetrabenazine.

A summary of the studies that met all inclusion criteria is presented in Table 1. Sample size, drug dose, duration of treatment and assessment tools varied greatly across studies. No head-to-head studies were available that compared the two interventions.

The tetrabenazine studies included one double-blind, crossover study (n = 12) and seven open-label studies (Total n = 381; range, n = 17–149). Tetrabenazine doses ranged from 6.25 to 300 mg/day (mean doses: 37.5–175 mg/day), administered one- to three-times daily. Tetrabenazine was generally well tolerated and effective in the treatment of TD. The most common side effects were parkinsonism, drowsiness/fatigue and depression, and authors of the studies generally agreed that tetrabenazine side effects were dose related and reversible.

Two open-label studies of tetrabenazine were conducted solely in TD patients. The most recent of these was a study by Ondo et al. [36] that included 20 patients with TD and used the Abnormal Involuntary Movement Scale (AIMS) as an outcome measure. In addition to TD, 45% of patients also showed mild evidence of parkinsonism, and 25% had akathisia at baseline. Participants were required to stop DRBA medications and other TD treatments for ≥30 days prior to starting tetrabenazine (mean dose: 57.9 mg/day; mean treatment duration: 20.3 weeks). Participants in this study were diagnosed with a psychiatric disorder or symptoms (unspecified psychosis, schizophrenia, bipolar disorder, agitation), gastrointestinal disorder or organic brain disorder. Cessation of antipsychotic medications is often not practical in patients with chronic psychotic disorders due to the risk of relapse, and it may have confounded treatment results due to unmasking or worsening of existing TD. TD severity was assessed by a single-blinded investigator who rated videos using the standardized AIMS, both at baseline and at approximately 3 months after commencing treatment. Significant improvement (p < 0.001) on the motor section of the AIMS was demonstrated at the end of treatment versus baseline values; no patient had unchanged or worsened TD (Table 2). The most common adverse events (≥10%) were sedation and parkinsonism (Table 3).

The second study that included only TD patients was a case series by Watson et al. [33] that included 23 patients with TD who were treated with tetrabenazine (mean dose: 91.3 mg/day; mean treatment duration not reported). Severity of involuntary movements was evaluated in three regions (face/mouth/tongue, trunk, limbs) using a 5-point involuntary movement scale (0 = none, 1 = minimal, 2 = mild, 3 = moderate, 4 = severe) and video recordings made for each patient before and during treatment. All patients in this study showed improvement with tetrabenazine treatment (range: 1–4 points) (Table 2). Side effects were minimal (all ≤10%), and the most common events were drooling and parkinsonism with tremor (Table 3).

The remaining studies were conducted in patients with various movement disorders, including TD. Among them was a crossover study by Asher and Aminoff [31] that enrolled patients with a diagnosis of TD (n = 12), Huntington's disease or dystonia. The specific psychiatric diagnoses of participants were not reported, but stable doses of antipsychotics and other medication were allowed throughout. In TD patients, the tetrabenazine mean dose was 175 mg/day. Study participants were informed of the crossover design, but were not told whether they started with tetrabenazine or placebo (although all were started on tetrabenazine) and when the crossover occurred. Participants were videotaped at baseline, 3 weeks (end of tetrabenazine treatment) and 6 weeks (end of placebo crossover); videotapes were reviewed by two blinded investigators who rated TD severity as slight, moderate or marked. The study used a 5-category rating system (marked, moderate, slight, none, worse) to assess treatment response (Table 2). Clinical significance of tetrabenazine treatment was based on TD severity being improved relative to both baseline and placebo treatment. Six of 12 patients showed marked or moderate improvement (50%). Tetrabenazine did not produce behavioral changes in participants with TD, and alteration of concomitant antipsychotics was not needed. Across all participants, the most common side effects (≥10%) were drowsiness, drooling and parkinsonism (Table 3).

Three open-label studies, conducted at a single site, investigated the effects of long-term tetrabenazine treatment on patients with TD, dystonia, Huntington's disease, Tourette's syndrome, myoclonus, tics and other movement disorders. The use of concomitant medications and specific psychiatric disorders were not reported in these studies. Response to tetrabenazine was scored approximately every 3–6 months by a single examiner, using a 5-point Likert scale to rate abnormal movements (1 = marked improvement, 2 = moderate improvement, 3 = mild to moderate improvement, 4 = poor or no improvement, 5 = worsening). In the 1988 study [34], 44 patients with TD received tetrabenazine 25–100 mg/day, with a mean treatment duration of 21.1 months. The majority of patients showed moderate to marked improvement of TD (Table 2). In the 1997 study [35], which appears to have included patients from the 1988 study who were still receiving treatment, the mean maximum tetrabenazine dose in patients with TD (n = 93) was 96.9 mg/kg and the mean treatment duration was 35.4 months. The percentage of patients who had moderate to marked improvement in this study was similar to the earlier study (93 and 71%, respectively; Table 2). The 2007 retrospective chart review [38] did not include patients who commenced treatment prior to 1997, thus excluding patients from the previous reports. The mean dose was 60.4 mg/day and mean duration was 2.3 years (all movement disorder patients, n = 490). Response rates did not vary over time, with the majority of TD patients showing moderate to marked improvement both after initiating treatment, and after long-term treatment with tetrabenazine (Table 2). Adverse events that occurred in ≥10% of tetrabenazine-treated patients (all movement disorders) in these studies were parkinsonism, drowsiness/fatigue, depression, insomnia and nervousness/anxiety (Table 3). Up to 23% of patients discontinued tetrabenazine treatment due to adverse events.

The tetrabenazine studies also included retrospective chart reviews by Paleacu et al. [37] and Miguel et al. [39]. Both studies examined the effects of tetrabenazine on TD and other movement disorders in patients who did not respond to other medications. The use of concomitant medications and specific psychiatric diagnoses were not reported in either study. In the chart review by Paleacu et al. [37], mean treatment duration was 22 months and mean dose was 76.2 mg/day (all movement disorder patients, n = 188). Improvement of movement disorders was rated using a Clinical Global Impression of Change scale, with scores ranging from -3 (marked worsening) to +3 (marked improvement). Approximately 60% of patients included in this chart review showed any improvement (mild, moderate and marked) of TD (Table 2). Side effects included somnolence, weakness, apathy and parkinsonism, but none of the reported side effects occurred in >10% of all patients (Table 3). In the chart review by Miguel et al. [39], mean treatment duration was 40 months and mean maximum daily dose was 37.5 mg/day (TD patients, n = 35). Clinical response was based on comparison of baseline with follow-up visits using a 3-category rating system (improved and asymptomatic, improved but symptomatic, poor or no clinical response). Charts showed that the majority of patients with TD exhibited clinically meaningful improvement based on assessment of any improvement with treatment (Table 2). The most common adverse events (≥10%) were parkinsonism, depression, anxiety and worsening of pre-existing psychiatric disorder (Table 3).

In addition to examining the effects of tetrabenazine on TD and other movement disorders, Kenney et al. [38] and Miguel et al. [39] reported the effects of treatment cessation. Both studies showed that movement disorder severity reverted when tetrabenazine was interrupted for as little as 1 week or discontinued permanently.

The valbenazine studies included: two randomized, double-blind, placebo-controlled trials, both with 6-week treatment duration; and one long-term, uncontrolled, extension trial with treatment of up to 48 weeks (Total n = 327). Valbenazine was administered once daily in all studies. In the Phase II study (KINECT 2; n = 100) [29], doses ranged from 25 to 75 mg/day and 76% of subjects reached maximum dose. In the Phase III study (KINECT 3; n = 227) [26] and its long-term extension (n = 198) [30], valbenazine-treated patients received 40 or 80 mg/day.

The valbenazine studies included adults with TD and stable schizophrenia/schizoaffective disorder or mood disorder. Stable regimens of concomitant medications for management of medical and psychiatric disorders were allowed. Changes in TD were assessed by consensus between two centralized video raters using standardized AIMS [40] assessments at baseline and at each study visit. Central AIMS video raters were blinded to treatment and to study visit.

Change in the AIMS total score (items 1–7) from baseline to week 6 was the primary end point for KINECT 2 and KINECT 3, and both studies showed significant improvement in ratings of severity of TD with valbenazine compared with placebo (p < 0.001; Table 2). Least squares mean differences between valbenazine and placebo were 2.4 in KINECT 2 and 3.1 (with the 80 mg/day dose) in KINECT 3. The long-term KINECT 3 extension study showed sustained improvement in the AIMS total scores through the end of valbenazine treatment (week 48), but TD reverted toward baseline levels of severity 4 weeks after treatment withdrawal at the end of the trial (week 52) (Table 2). The percentages of participants achieving a rigorous AIMS response threshold, defined as a ≥50% total score improvement from baseline, were significantly higher with valbenazine compared with placebo in both studies (KINECT 2, 49 vs 18%; KINECT 3, 40% [80 mg/day] and 24% [40 mg/day] vs 9%; all p < 0.05).

The Clinical Global Impression of Change-TD (CGI-TD) [40], which was scored by an onsite investigator who was blinded to treatment, indicated clinically meaningful but not statistically significant improvements with valbenazine compared with placebo in KINECT 3 (Table 2). In the KINECT 2 study, however, results for CGI-TD response, defined as a rating of ‘much improved’ or ‘very much improved’ (score ≤ 2), were significantly greater for valbenazine compared with placebo (p < 0.001), and significantly more patients showed a response with valbenazine (67 vs. 16%, p < 0.0001).

Similarly, in the KINECT 3 extension study, improvements in scores from baseline through week 48 were maintained in AIMS total and the CGI-TD scores (Table 2). The proportion of patients maintaining a response was sustained based on both the CGI-TD scores (76% [80 mg/day]; 59% [40 mg/day]) and ≥50% improvement in AIMS scores (52% [80 mg/day]; 28% [40 mg/day]).

Valbenazine was generally well tolerated, especially considering that most participants were taking a concomitant antipsychotic medication (KINECT 2, ≥40% of patients; KINECT 3, 86% of patients) [26,29]. The most common adverse events that were reported with valbenazine were fatigue, headache, somnolence and reduced appetite (all ≤10%) (Table 3). In the long-term extension study, headache and urinary tract infection (7% each) were the most common adverse events. Psychiatric status remained stable throughout valbenazine treatment and no other clinically relevant safety signals were observed.

As is often the case with systematic reviews [44], one of the main limitations of this analysis was the paucity of details available in some publications, especially the observational tetrabenazine studies. Moreover, as discussed earlier, the heterogeneity of the studies precluded the possibility of conducting a more formal meta-analysis. Therefore, only descriptive summaries could be provided and any interpretation of the results should be considered qualitative in nature. Although purposefully broad search terms were used to capture as many potential articles of interest within PubMed and Embase (i.e., the most frequently used literature databases in the USA), it is possible that this review did not include relevant publications that were only indexed in other databases. However, it seems reasonable to assume that the publications included in this review are adequately representative.