Long-Term Efficacy and Safety of Botulinum Toxin Injections in Dystonia

Local chemodenervation with botulinum toxin (BoNT) injections to relax abnormally contracting muscles has been shown to be an effective and well-tolerated treatment in a variety of movement disorders and other neurological and non-neurological disorders. Despite almost 30 years of therapeutic use, there are only few studies of patients treated with BoNT injections over long period of time. These published data clearly support the conclusion that BoNT not only provides safe and effective symptomatic relief of dystonia but also long-term benefit and possibly even favorably modifying the natural history of this disease. The adverse events associated with chronic, periodic exposure to BoNT injections are generally minor and self-limiting. With the chronic use of BoNT and an expanding list of therapeutic indications, there is a need to carefully examine the existing data on the long-term efficacy and safety of BoNT. In this review we will highlight some of the aspects of long-term effects of BoNT, including efficacy, safety, and immunogenicity.


Introduction
BoNT is the most potent biologic toxin found in nature. Its pharmacological action results in inhibition of acetylcholine release into the neuromuscular junction. As a result of local chemodenervation BoNT relaxes abnormally contracting muscles which has been shown to be an OPEN ACCESS effective treatment in a variety of movement disorders and other neurological and non-neurological disorders. After several months of deliberation and a requirement for various precautions, including the presence of resuscitation crash cart, the Baylor College of Medicine Institutional Review Board for Human Research first granted approval for the use of BoNT to one of the authors (JJ) who injected the first patient, a 45 year old woman with blepharospasm, on 12/15/1981. Since the first double-blind, placebo-controlled trial demonstrated efficacy of BoNT-A in cervical dystonia, published in 1986 [1], and in cranial-cervical dystonia, published in 1987 [2], BoNT has been used to treat hundreds of different conditions not only in Neurology but nearly all fields of Medicine. The results from the initial trial in cranial-cervical dystonia at Baylor College of Medicine [2], blepharospasm, a form of focal dystonia, was the first indication approved by the United States Food and Drug Administration (FDA) for the treatment with BoNT-A in 1989 [3].
Currently there are four BoNT products commercially available in the US approved for various therapeutic and cosmetic indications with the following non-proprietary and brand names: OnabotulinumtoxinA (Botox ® [Type A]); IncobotulinumtoxinA (Xeomin ® [Type A]); AbobotulinumtoxinA (Dysport ® [Type A]); RimabotulinumtoxinB (Myobloc ® [Type B]) [4]. In addition to FDA-approved indications, there is an expanding list of off-label therapeutic uses [5][6][7][8][9]. In this review we will highlight some of the aspects of long-term effects of BoNT, including efficacy, adverse effects, and immunogenicity.

Long-Term Experience and Efficacy
Because it would not be feasible or ethical to conduct long-term double-blind, placebo-controlled studies, data from published longitudinal studies are limited by lack of objective assessments and relatively small sample size. Only a few studies have been published regarding the long-term effects of BoNT therapy for the treatment of dystonia over five years or longer (Table 1). In 2005 we published longitudinal follow-up data on 45 patients who had received BoNT treatments continuously for at least 12 years at Baylor College of Medicine Movement Disorders Clinic [10]. Patients were followed for a mean of 32 visits and a mean of 16 years. The most frequent reasons for treatment were cervical dystonia (43%), cranial dystonia (26%), and blepharospasm (11%). Patients' response was rated according to a 0-4 scale (0 no effect to 4 marked improvement in severity and function). Over time, there was no significant change in latency and total duration of response to treatment and the dose per visit, peak duration of response, global rating, and peak effect showed incremental improvement over time. One third of patients developed side effects after their initial visit compared with 22% at their most recent visit. The most common side effects were dysphagia and ptosis. Antibody testing by the mouse protection assay (MPA) was performed in 22 patients due to non-responsiveness; blocking antibodies were confirmed in 4 (18%) of these patients. Sixteen of the antibody-negative patients resumed responsiveness by dose adjustment, whereas 2 persisted as non-responders. These findings suggest that BoNT is an effective long-term treatment with adequate and persistent therapeutic response and mild side effect profile. The relatively high frequency of immunoresistance was due to the use of original OnabotulinumtoxinA, which had a relatively high protein load (see below).
A retrospective analysis of the long term efficacy of BoNT-A over a 10-year period in 235 subjects with cervical dystonia, hemifacial spasm, blepharospasm, and other focal/segmental dystonia found that at 2 years the highest response rate of sustained benefit (defined as continued improvement of 50% or more from baseline) was in patients with hemifacial spasm (96%), followed by patients with blepharospasm (92%) and with cervical dystonia (68%) [11]; this was similar at 5 years. The BoNT doses used during the first year were compared with doses used during second to fifth year, and doses used after the fifth year. Minor side effects developed in 27% of the patients at any one time, occurring in 4.5% of treatment cycles. Patient satisfaction increased after 5 years of treatment with an average reported benefit of 75.8%.
Another retrospective review of 20 patients with focal hand dystonia followed for 10 years, up to 16 years showed mild average benefit in 11 subjects with a trend towards larger benefit in women [12]. This cohort required a gradual increase in dose over time, patients received a higher mean dose at the end of the follow-up period compared to the initial treatment (50 vs. 25 U respectively, p < 0.00005), which could partly be explained by a relatively low initial dose. Focal hand dystonia has a lower overall response rate, with about 50% of patients receiving at least mild benefit compared to 80% for cervical dystonia [13] and over 90% for blepharospasm [14]. More than one third of musicians reported long-term benefit and improvement in their performance ability in a retrospective review [15] of 84 (74 men) subjects with focal task-specific dystonia by estimating their playing ability compared to their skill level prior to onset of dystonic symptoms before and during EMG-guided BoNT treatment and rated their cumulative treatment response using a self-rating scale.
Post-hoc analysis of two prospective, long-term, open-label extension studies of AbobotulinumtoxinA in cervical dystonia [16] demonstrated that treatment was generally well tolerated with no major differences in the incidence of adverse events. In both studies, patients were randomized to receive a single treatment cycle of AbobotulinumtoxinA 500 U diluted in 1 mL preservative-free normal saline or placebo. A total of 137 subjects entered the first double-blind, placebo-controlled study [17] and 131 received at least one treatment cycle of AbobotulinumtoxinA. A total of 116 subjects entered the second double-blind, placebo-controlled study [18], 108 of whom entered the open-label study. Based on the analysis of the combined cohort of 239 patients who entered the two-extension, open-label, studies all treatment cycles resulted in improvements in mean Toronto Western Spasmodic Torticollis Rating Scale (TWSTRS) total scores 4 weeks post-injection. The mean baseline TWSTRS scores at the time of the first cycle were 44.5 ± 8.8 and 42.5 ± 10.2 for the first and second studies, respectively; and the mean TWISTRS score changes from baseline were −13.7 ± 10.9 and −16.2 ± 11.5 for the first and second studies, respectively. At the time of the third cycle, the baseline scores were 41.0 ± 10.7 and 32.75 ± 11.7 for the first and second studies, respectively; and the mean changes were −12.2 ± 10.3 and −10.8 ± 13.0 for the first and second studies, respectively. However, increasing the dose of AbobotulinumtoxinA above the initial 500 U dose was not observed to result in an incremental improvement in response as measured by the TWSTRS.
A survey of 155 patients with cervical dystonia treated with BoNT-A injections over a period of 6 years [19], 133 (86.6%) individuals completed the survey; 78% (104) continued and 22% (29) stopped therapy. Two thirds of the patients who continued BoNT-A treatments reported the injections always helped, whereas one quarter estimated one set of injections did not help. One third of those continuing treatment reported the first injection was most helpful, whereas another one third felt all injections were similarly effective. After an initial adjustment, BoNT-A dosages and frequency of treatment remained stable.
A report of efficacy of BoNT-A in 78 patients with idiopathic cervical dystonia treated for a median of 5.5 years (range 1.5-10 years), as evaluated by the patients themselves and their neurologists, showed a high degree of patient satisfaction [20]. The effect of BoNT treatment was scored on a seven-point scale (0 excellent to 6 worse), being excellent, good or moderate in 85%. The independent scores of the treating neurologists were excellent, good or moderate in 92%, and correlated well with the patients' scores. There was also a marked reduction of 'Global Burden of Disease', as expressed on the Visual Analog Scale (0 designating no burden and 10 the worse imaginable burden). By combining these outcome measures into two major groups, 67% of the patients were designated as having a good effect, and 33% had an unsatisfactory effect.
The published data provides evidence that BoNT products not only exert safe and effective symptomatic relief but also long-term benefits, favorably modifying the natural history of this disease, such as lower risk of contractures in patients with cervical dystonia [25]. The adverse events associated with chronic, periodic exposure to BoNT type A or B injections are generally minor and self-limiting, and even decrease over time [26]. The presence of neutralizing antibodies against the toxin is only responsible for a minor proportion of the small group of dystonia patients that develop persistent therapeutic resistance.
The following Table summarizes the most relevant findings from published studies of efficacy, safety, and side effect profile of long-term BoNT therapy. The included studies were selected based on the main treating diagnosis of any type of dystonia with the periodic use of BoNT injections with a mean follow-up period of at least 5 years, including description of either total or per-session mean dose of the toxin, outcome measures of response/effectiveness, and report of the adverse events. If a study did not have at least a range follow-up period of 60 months, it was omitted from this table.

Immunogenicity and Therapeutic Response
Lack of response to BoNT treatment for dystonia and other movement disorders may occur for myriad reasons including inadequate dose, inappropriate muscle selection, concomitant drug therapy, dynamic disease change, and the development of neutralizing antibodies [10,49,50]. As a foreign protein, BoNT is inherently immunogenic, but because it is administered in only extremely small quantities at relatively long inter-dose intervals, most patients do not develop immunoresistance to BoNT. Nevertheless, BoNT elicits antibodies in some patients that can neutralize the neurotoxin, as a result of which the patient becomes unresponsive that type of BoNT [5]. The presence of neutralizing antibodies that act by blocking binding of the BoNT complex to the presynaptic membrane is one explanation for the occasional reports of lack of response to BoNT injections [51].
Although immunogenicity was an important consideration in the past [52], in today's clinical practice poor or no response to BoNT is more frequently due to factors other than immunogenicity [29,49,53,54]. A prospective, open-label, multicenter study of 326 patients with cervical dystonia evaluated the immunogenicity of BoNT-A [49]. Seventy-seven percent subjects completed the study, receiving a mean dose per session ranging from 148.4 to 213.0 U and an overall mean cumulative dose of 1576 U over a mean of 2.5 years. The two main reasons for discontinuation of BoNT treatment were loss to follow-up (28%) and adverse events (17%). Four of 326 (1.2%) subjects developed neutralizing antibodies as measured by the MPA, one continued to respond to treatments, and the other three lost clinical responsiveness. Two of the 3 patients had a clinical resistance test (test injection of 20 U of BoNT-A placed unilaterally into the frontalis) performed and only one was found to indicate unresponsiveness. Although the MPA is considered the "gold standard" with relatively high specificity, it may lack sensitivity in identifying those with low antibody titers, thus underestimating the number of subjects with immunoresistance [55].
The previous presence of neutralizing antibodies may not be a predictor of response to BoNT therapy. A randomized, multicenter, double-blind study evaluated the efficacy of OnabotulinumtoxinA versus placebo for the treatment of cervical dystonia. The impact of the subjects' neutralizing antibody status at study entry on clinical response was analyzed. A total of 214 subjects were enrolled in period 1 and received treatment with OnabotulinumtoxinA at a mean dose of 241.2 U (range, 95-360 U). One hundred seventy subjects entered period 2 and were randomized to receive Onabotulinumtoxin A (n = 88) or placebo (n = 82). Of the 114 subjects with valid samples at both study entry and exit, 19 subjects showed the presence of neutralizing antibodies at study exit; however, 17 of these samples tested positive at study entry, indicating that only 2 (2%) of 97 subjects seroconverted from negative to positive over the course of 2 treatments in this 22-to 26-week study. The two subjects remained responsive to OnabotulinumtoxinA during both the open and blinded treatment periods. OnabotulinumtoxinA improved disabling features associated with cervical dystonia, with a robust response in both groups compared to baseline [53].
BoNT complexing proteins elicit antibodies in about half of patients [56,57], but only a small proportion of induced antibodies are actually neutralizing (blocking) antibodies. Furthermore, there is evidence suggesting that the anamnestic immunologic response to BoNT can wane with time, but can be reactivated by repeat exposure to the same type of BoNT. BoNT positive antibody status can revert to negative status over time. Seven patients, with a mean age 56 years, six with cervical dystonia and one with oromandibular dystonia, immunoresistant to BoNT-A therapy with positive antibodies, were retested by MPA for the persistence of neutralizing antibodies [58]. The mean duration of symptoms was 197 months and received a total mean cumulative dose of 1659 U. All patients reverted from BoNT-A positive antibody status to negative. Six antibody positive patients were reinjected with BoNT-A after they reverted to negative antibody status. All improved after the reinjections, but three lost their response to subsequent injections, which was confirmed by re-emergence of antibodies. In four of these patients, response to repeat BoNT-A injections was confirmed by weakness of the right frontalis muscle injected simultaneously. This unique group of patients responded favorably to repeat BoNT-A injections, but some lost the benefit with subsequent injections.
Differences among BoNT preparations, both within and between serotypes, lead to differences in clinical performance, safety, antigenicity, and specificity [59]. Although recently challenged [54], RimabotulinumtoxinB (Myobloc ® [Type B]) has been reported to have a high rate of immunogenicity [60,61]. This high frequency of antibody formation may be possibly related to a large amount of protein in the BoNT-B product. IncobotulinumtoxinA (Xeomin ® [Type A]); is a preparation of BoNT type A that is structurally free of NAPs with high specific biological activity [62], but it is not yet clear whether this property translates into lower antigenicity [63].

Adverse Effects and Treatment Failures
Despite the widespread commercial acceptance of BoNT for cervical dystonia, there is lack of adequate follow-up data to determine the reasons why patients stop therapy. In a survey of 155 patients with cervical dystonia who were treated over 6 years with BoNT of 133 (86.6%) individuals who returned the surveys, 104 continued and 29 stopped therapy. From the latter group, 11 individuals had only received one or two injections. Prior surgical treatment for cervical dystonia did not influence their decision to stop therapy [19].
Weakness in cervical musculature and dysphagia are the most common BoNT treatment-related side effects in patients with CD [19,29,64]. Dropout from treatment can also be secondary to symptom improvement, remission, or location/access-related issues [11,29,34]. In one study involving a total of 616 patients [29], 155 of whom were lost to follow-up, there was a total of 173 reasons for discontinuation. A large group of those patients (48) had not actually discontinued BoNT-A therapy but had merely switched to other treatment centers. The corrected overall dropout rate was thus 20% (126/616 patients). The common reasons (number of patients/percentage) for withdrawing were primary nonresponse (33/26.1%), secondary nonresponse (17/13.4%), adverse events (27/21.4%), remission/improvement (26/20.6%), and other (34/26.9%) which included inconvenience of cost, travel, relocation, discontinuation by physician, lost to follow-up, and unknown.
In a similar 10-year retrospective analysis of 235 clinic patients with various movement disorders treated with BoNT-A, 28% of patients discontinued treatment during the follow-up period due to a variety of reasons [11]. Overall, 9.1% (21 patients) had primary resistance, which is highly surprising as in our practice we almost never see primary nonresponders unless they were previously vaccinated against BoNT, as is the case in some military personnel. Apparently, 7.5% 18 patients) developed secondary resistance, but it is not clear who this was verified and no MPA or other assays were apparently performed. Only 1.3% of patients discontinued therapy due to intolerable adverse effects; discontinuation for any reason was also low after 5 years. The main reason for stopping treatment was lack of benefit which may be secondary to inadequate muscle selection or injection technique, evolution of primary disease, or development of neutralizing antibodies. Another long-term cohort study of 100 cervical dystonia patients treated with AbobotulinumtoxinA followed for an average of 5 years [34] showed high adherence to BoNT treatment over time. More than 60% (57 of 90) of these patients continued with BoNT injections after up to 12 years. The most common reasons for stopping treatment were inconvenience (long travel distances, costs) 36% (12 of 33), side effects 33% (11 of 33), and significant improvement of dystonia 15% (5 of 33). Other reasons for discontinuation were secondary nonresponse (3), primary nonresponse (1), and complete remission (1).

Conclusions and Future Directions
This review of longitudinal data provides evidence of long-term efficacy and safety of BoNT in the treatment of dystonia. In addition, chronic BoNT treatment has changed the natural history of dystonia. For example, contractures, which often occurred as a consequence of prolonged abnormal posture, are now rarely seen due to prevention of this dystonia-related complication by early institution of BoNT therapy. We chose to focus on dystonia as this is the earliest approved indication [2,65], but similar longitudinal experience is accumulating in other clinical conditions chronically treated with BoNT.
With the development of new BoNT preparations that seem to have lower risk for antigenicity, immunoresistance appears to be a less important cause of loss of response than in the past. Further longitudinal data are needed to determine which rating scales and quality of life outcome measures are best suited for long-term follow-up studies and to better understand the reasons for withdrawal from continued BoNT therapy.
Relative advantages and disadvantages of the various BoNT products need to be addressed not only in short-term, controlled studies but in well-designed long-term, observational studies. Although currently available BoNT drugs are safe and effective, the research and development of new products should focus not only on issues related with antigenicity, but also on selectivity based on duration of action, diffusion properties, and novel delivery routes. Future studies should also evaluate the most appropriate method of muscle selection for BoNT injection and techniques to prolong its effects.
Finally, there is a need to study the long-term effects of BoNT on central neurophysiologic changes and neuroplasticity. The notion of central remodeling of cortical motor maps as a result of reduced afferent input following BoNT [66] needs to be further explored as a potential neurophysiologic and clinical consequence of long-term BoNT therapy.