Assessment: Botulinum neurotoxin for the treatment of movement disorders (an evidence-based review)

Blepharospasm.

Blepharospasm is a focal dystonia characterized by involuntary contraction of orbicularis oculi, causing involuntary closure of the eyes. Prior to BoNT, there were no effective medical or surgical treatments for this disorder. Blepharospasm was one of the first studied indications for BoNT treatment. The open label observations were felt to be so dramatic that there have been only a few attempts to perform properly controlled clinical trials. In 1989, Botox® received Food and Drug Administration (FDA) approval for blepharospasm, including benign essential blepharospasm or VII nerve disorders, in patients 12 years and older. Two efficacy trials fulfilled criteria for Class II (table e-1 on the Neurology® Web site at www.neurology.org).

One Class II trial was a double-blind comparison of injecting BoNT-A into one eyelid and saline into the other.³ Six patients received BoNT-A (Botox®) at a dose of 20 U/eye for the active treatment. Blinded rating of videotapes showed bilateral reduction in blepharospasm that was greater on the side injected with active toxin. The other Class II study was a double-blind, prospective, crossover trial of 11 patients using BoNT-A (Botox®) at 25 U/eye, supplementing up to 50 U/eye 1 month later if needed.⁴ Assessment by physician observation and rating of videotapes, using the Fahn scale and by patient subjective rating, showed significant efficacy lasting a mean of 2.5 months. Adverse effects were generally mild (blurred vision, tearing, ptosis, and ecchymosis).

One Class II and one Class III study compared two different brands of BoNT-A (Botox® and Dysport®). In the Class II study, there were 212 patients evaluated in a crossover design using a 4:1 dose ratio of Dysport® to Botox®.⁵ The primary endpoint, duration of effect, was similar for the two products. The Class III study used a parallel design of 42 patients without blinded raters and also used a dose ratio of 4:1.⁶ Duration of action was again the primary endpoint, and this endpoint and others including number of booster doses needed, latency of effect, clinical efficacy, and adverse reactions were similar for the two products. A Class I study compared Xeomin® and Botox®, using equivalent doses in 300 patients, with 256 patients completing the study. There was no difference in efficacy or adverse effects between the two formulations.⁷

Hemifacial spasm.

Hemifacial spasm is characterized by a combination of unilateral clonic and tonic spasms of the muscles innervated by the facial nerve. Treatment options include oral pharmacologic therapies, including carbamazepine, baclofen, and benzodiazepine, and resulting in limited efficacy, and microvascular decompression of the facial nerve, a highly invasive procedure. Encompassed in the category of VII nerve disorders, hemifacial spasm is FDA approved. One efficacy trial of BoNT fulfilled criteria for Class II, and one for Class III (table e-2). The Class II study⁸ of 11 patients was a prospective, blinded trial with four arms: an arbitrary dose based on clinical experience of between 2.5 and 10 units of BoNT-A (Botox®), half the dose, double the dose, and saline placebo. Each subject cycled through the four treatment arms in a random order. Using a clinical scale to rate videotapes and a patient subjective scale, 84% had objective improvement with at least one of the active doses with a trend for better response with higher dose; only one patient improved on placebo. Seventy-nine percent reported subjective benefit lasting a mean of 2.8 months with active therapy. Weakness of the face, generally mild, was the most common adverse effect (97%). Other adverse effects included bruising, diplopia, ptosis, and headache.

A Class III study⁹ was a double-blind, prospective, parallel design study with only four patients per group using individualized therapy (dose range 2.5 to 40 units) with BoNT-A (Botox®) in the active arm. Ninety-three patients studied in an open label fashion were also reported. There was greater improvement on a clinical scale with BoNT than with a saline placebo. Benefit lasted a mean of 3.8 months. Side effects, present in 63% of patients, were generally mild and included dry eye, mouth droop, and ptosis.

One Class II study compared Botox® and Dysport® in a parallel design without placebo control or blinded raters.⁶ There was a dose ratio of 4:1 for Dysport® to Botox.® The primary endpoint (duration of action) and other endpoints (number of booster doses needed, latency of effect, clinical efficacy, and frequency of adverse reactions) were similar for the two products. Benefit lasted 2.6–3.0 months.

Cervical dystonia.

Cervical dystonia (CD) is a focal dystonia causing involuntary activation of the muscles of the neck and shoulders resulting in abnormal, sustained, and painful posturing of the head, neck, and shoulders. There are limited data assessing oral medications for cervical dystonia. Recent surgical studies, including deep brain stimulation, show promise. Out of approximately 80 studies of BoNT in the treatment of CD, 14 controlled studies were identified, including seven Class I studies (four with BoNT-A, three with BoNT-B) (table e-3).^10–16 Botox® and Myobloc® are FDA approved for use in CD.

Three Class I studies enrolled BoNT-naïve CD subjects. One study¹⁶ evaluated 55 subjects over 12 weeks. Subjects were stratified by primary type of torticollis and randomized to BoNT-A or placebo. Maximal benefit occurred at 6 weeks following injection with improvement in functional capacity, head turning, pain, and subjective assessment. Adverse events included dysphagia, neck weakness, and fatigue.

A second Class I study compared low, intermediate, and high doses of BoNT-A to placebo in 75 BoNT-naive CD subjects with rotational torticollis.¹² At 1 month, the intermediate- and high-dose groups demonstrated improvement compared to placebo (p < 0.05). Subjective improvement occurred in 45–50% and was significant at 1 and 2 months for the high-dose group, and at 2 months for the intermediate-dose group. More adverse events occurred with higher doses, including neck weakness, voice changes, and dysphagia.

A third Class I study in BoNT-naïve subjects compared BoNT-A and trihexyphenidyl (mean dose 16.25 mg; range 4–24 mg) at 3 months following treatment.¹⁵ BoNT-A was superior to trihexyphenidyl for TWSTRS disability (2 points), Tsui scale (5 points), and general health perception (6 points). Although there was greater improvement in TWSTRS pain score with BoNT (2 points), this did not reach statistical significance. The total TWSTRS and TWSTRS severity scores were not given for either group. The trihexyphenidyl group had more adverse events (76 events vs 31 for BoNT-A, p < 0.0001).

Four Class I studies enrolled subjects with previous response to BoNT. Three studies assessed safety and efficacy of BoNT-B,^10,11,14 and one assessed safety and efficacy of BoNT-A.¹³ One of these studies randomized 109 subjects with CD to placebo, medium dose, or high dose of BoNT-B.¹⁰ There was an improvement in total TWSTRS scores at 1 month for both treatment groups compared to placebo (medium dose p = 0.01; high dose p = 0.01), with return to baseline by 3 months. The TWSTRS severity and pain subscales and patient and physician global scales showed similar benefit. Adverse events were greater in the BoNT-B treated groups, with dry mouth and pain occurring in a dose-dependent way. All adverse events were mild.

A similar study assessed the effect of BoNT-B compared to placebo in 77 patients with CD who developed resistance to BoNT-A.¹¹ At 1 month following injection, the BoNT-B group had more improvement in total TWSTRS score (21% vs 4% in placebo, p = 0.0001). Treatment with BoNT-B improved the TWSTRS severity, disability, and pain subscales, and physician and patient global scores. Dry mouth occurred in 3% of placebo subjects and 44% of the BoNT-B group. A study evaluated BoNT-A compared to placebo in 80 subjects with CD previously treated with BoNT-A.¹³ This study showed improvement in TWSTRS total score and each of the subscales of the TWSTRS for severity, disability, and pain at 1 month with only blurred vision and neck weakness occurring more frequently than placebo.

Focal limb dystonia.

Most studies of BoNT in focal limb dystonia deal with the upper extremity. Although no controlled trials of BoNT exist for lower limb dystonia, some larger studies do include these patients. The term “focal hand dystonia” is used here to encompass writer's cramp, other occupational hand dystonia, and nontask-specific hand dystonia. The pattern of limb dystonia varies widely among patients. There are no effective alternative medical or well-established surgical therapies for focal limb dystonia. The use of BoNT to treat limb dystonia requires thoughtful technique including customization of doses and muscle selection.

There is Class I and Class II evidence for focal limb dystonia¹⁷ (table e-4). The Class I trial randomized 40 patients with writer's cramp in a double-blind design to BoNT or an equivalent volume of saline placebo. Injected muscles were chosen based on clinical examination. Participants with inadequate or no response were offered a second injection 1 month later. The primary outcome measure was the subject's stated desire to continue injection. Seventy percent of those randomized to BoNT wished to continue treatment compared to 32% of those receiving placebo (p = 0.03). Significant improvement was also found in BoNT-injected subjects compared to those receiving placebo in secondary outcome measures including a visual analog scale, symptoms severity scale, writer's cramp rating scale, and assessment of writing speed, but not in the functional status scale. Temporary weakness and pain at the injection site were the only adverse events reported.

One Class II trial¹⁸ was a prospective, double-blind, crossover study of 17 patients with several forms of limb dystonia, including lower extremity (3 patients) and secondary dystonia (4 patients). Subjects received a series of four injections in random order, one with a dose of BoNT that the investigators judged to be “optimal,” one at half the optimal dose, one at double the dose, and one with saline placebo. Using a patient subjective scale, 82% of patients receiving BoNT had benefit compared to 6% (one patient) who received placebo. Using physician rating of videotapes, 59% improved with active treatment and 38% with placebo (not significant). There was no dose-response relationship for benefit, and there was a large degree of interobserver variability. The authors attributed the lack of significance in physician ratings to an inadequate outcome evaluation. The main side effect was focal weakness that followed 53% of BoNT injections and was more likely at the higher doses. Weakness occurred with 13% of placebo injections. Other adverse effects included muscle stiffness, pain, and malaise.

Another Class II study used a placebo-controlled, double-blind, crossover design in 20 patients with writer's cramp.¹⁹ Muscle selection was guided by clinical examination; dose of BoNT-A was based on investigator experience. Outcome assessments included evaluation of writing speed, accuracy, writing samples, and patients' subjective report. There was significant improvement with BoNT therapy in the objective measures, but not in patients' own assessments. Focal weakness was the only adverse effect and was severe enough to worsen pen control in one patient. The authors noted that this study evaluated only the first active treatment session that a patient received, so that the benefit obtained was likely not optimal.

Another Class II trial was a double-blind, placebo-controlled, crossover in 10 patients with focal hand dystonia.²⁰ Muscles and BoNT-A doses were selected and optimized during a period of open treatment preceding the controlled study. Patient subjective rating and observer rating of videotapes during activities applicable to individual dystonia were the outcome measures. Eight patients had improved subjective rating and six had improved videotape rating with BoNT compared with placebo. Weakness was present in the injected muscles in 80% of subjects with active treatment.

Three Class II studies evaluated technical issues of BoNT administration (table e-4). In one study, a blinded, randomized, crossover design was used to compare continuous muscle activation to immobilization immediately after BoNT injection.²¹ Blinded evaluation of handgrip strength and writing revealed a significant increase in focal weakness with continuous muscle activity, but no subjective or objective improvement in writing. In another Class II study, patients were randomized to one of two muscle localization techniques: EMG recording or electrical stimulation.²² Injections guided by either technique were equally effective in producing weakness in the target muscle. The accuracy of muscle localization with and without EMG was evaluated in a third trial.²³ Only 37% of needle placements based on surface anatomy were localized in the targeted muscle.

Laryngeal dystonia.

Laryngeal dystonia (spasmodic dysphonia) generally presents as adductor type (ADSD) and less frequently as abductor type of spasmodic dysphonia (ABSD). ADSD is characterized by a “strain-strangle” voice, while ABSD produces a breathy and hypophonic voice. There are no effective alternative medical or surgical therapies for spasmodic dysphonia. There is one Class I study of BoNT²⁴ (n = 13) of patients with ADSD. This double-blind, randomized, parallel group study compared seven patients receiving BoNT with six receiving saline. Outcome measures included instrumental quantitative measures of voice function and patient ratings. Significant benefit was found in the BoNT-injected group (p = 0.01) (table e-5). One Class III study found that the addition of voice therapy following BoNT in ADSD prolonged benefit from BoNT treatment.²⁵ Another found that voice rest 30 minutes after BoNT injection prolonged the benefit of BoNT.²⁶ One Class III study of 15 patients with ABSD²⁷ did not find a significant difference using either percutaneous or endoscopic injection technique.

Tics.

Typically associated with Tourette syndrome, tics are relatively brief, intermittent movements (motor tics) or sounds (vocal or phonic tics), usually preceded by a premonitory sensation.²⁸ While antidopaminergic drugs (neuroleptics) are often effective in treating troublesome multifocal tics, these drugs often produce undesirable side effects, particularly in patients with focal tics, such as blinking, blepharospasm, head jerking, neck twisting, and loud vocalizations, including coprolalia. Although confined to a limited anatomic distribution, such focal tics may be a source of embarrassment and may result in functional blindness, local discomfort, and social isolation.

In initial open label Class IV studies, injections of BoNT in the muscles involved in the motor and phonic tics was associated with a moderate to marked reduction in the intensity and frequency of the tics, and nearly complete abolishment of the premonitory sensation. In a Class IV study of 35 patients treated in 115 sessions for troublesome or disabling tics, the mean peak effect response was 2.8 (range 0 = no effect; 4 = marked improvement in both severity and function).²⁹ The mean duration of benefit was 3.4 months (up to 10.5). Latency to onset of benefit was 3.8 days (up to 10). Twenty-one of 25 patients (84%) with notable premonitory sensory symptoms derived marked relief of these symptoms from BoNT (mean benefit: 70.6%).

In a Class II study of 18 patients with simple motor tics, there was a 39% reduction in the number of tics per minute within 2 weeks after injection with BoNT, as compared to a 6% increase in the placebo group (p = 0.004, table e-6).³⁰ In addition, there was a 0.46 reduction in “urge scores” with BoNT compared to a 0.49 increase in the placebo group (p = 0.02). This study lacked the power to show significant differences in other measured variables, such as severity score, tic suppression, pain, and patient global impression. The full effect of BoNT may not have been appreciated at 2 weeks. The study employed a single treatment session protocol that does not reflect the clinical practice of evaluating patients after several adjustments in doses and sites of injections. Furthermore, since subjects “did not rate themselves as significantly compromised by their treated tics,” it is likely that their symptoms were relatively mild at baseline.

Tremor.

Tremor, an oscillatory movement produced by alternating or synchronous contractions of antagonistic muscles, is the most common movement disorder. While propranolol and primidone usually ameliorate mild or moderate essential tremor, pharmacotherapy is usually not sufficient to control a high-amplitude tremor that impairs activities of daily living. In such cases of disabling tremor, local injection of BoNT may be used before considering more aggressive intervention such as thalamic deep brain stimulation.

A Class II placebo-controlled study evaluated 25 patients with hand tremor of 2 (moderate) to 4 (severe) on the tremor severity rating scale³¹ (table e-7). Subjects were randomized to receive either 50 units of BoNT-A (Botox®) or placebo injections into the wrist flexors and extensors of the dominant limb. If patients failed to respond to the initial injection, they were eligible to receive another injection of 100 units 4 weeks later.

Rest, postural, and kinetic tremor were evaluated at 2- to 4-week intervals over a 16-week study period, using tremor severity rating scales, accelerometry, and assessments of improvement and disability. There was significant improvement on the tremor severity rating scale 4 weeks after injection in patients treated with BoNT as compared to placebo, and this effect was maintained for the duration of the study. Four weeks after injection, 75% of BoNT-treated patients vs 27% of placebo-treated patients (p < 0.05) reported mild to moderate improvement. Functional rating scales did not improve although trends were observed for some items. Postural accelerometry measurements showed a 30% reduction in amplitude in 9 of 12 BoNT-treated subjects and in 1 of 9 placebo-treated subjects (p < 0.05). Although all patients treated with BoNT reported some degree of finger weakness, no severe, irreversible, or unexpected adverse events occurred.

There were similar results in another Class II multicenter, double-blind, controlled trial that utilized a similar protocol and involved 133 patients with essential tremor.³² The patients were randomized to receive 50 or 100 U of Botox® into wrist flexors and extensors and were followed for 4 months. The study showed significant improvement in postural tremor, but only minimal improvement in kinetic tremor and functional assessments.

The study design of both Class II studies limits their applicability to clinical practice. Both used a rigid treatment protocol that employed a fixed BoNT dose and a predetermined set of muscles. In practice, dosages and injected muscles are often individually chosen on the basis of tremor pattern.

An underpowered Class II study of 10 patients with head tremor³³ did not show a statistically significant benefit in BoNT-treated patients. There are two Class IV open-label studies in voice tremor^34,35 that showed modest improvement from baseline in objective acoustic and subjective measures after unilateral or bilateral BoNT injection.