Brain, Vol. 124, No. 4, 769-783,
April 2001
© 2001 Oxford University Press
Prospective study of selective peripheral denervation for botulinum-toxin resistant patients with cervical dystonia
1 University Departments of Clinical Neurology and 2 Neurosurgery, Institute of Neurology, National Hospital for Neurology and Neurosurgery, and 3 Reta Lila Weston Institute of Neurological Studies, Royal Free and University College Medical School, University College London, UK
Correspondence to:
Dr Kailash P. Bhatia, Department of Clinical Neurology, Institute of Neurology, Queen Square, London WC1N 3BG UK E-mail: k.bhatia{at}ion.ucl.ac.uk
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionConclusionsAppendix IReferences |
|---|
We have carried out a prospective study of selective peripheral denervation (SPD) in cervical dystonia (CD) patients with primary or secondary botulinum toxin (BT) treatment failure using independent standardized assessment. Patients referred for surgery had a standardized clinical examination, neck muscle EMG, videofluoroscopic swallow and CT of the cervical spine, and were selected for surgery on the basis of the results of these investigations. CD severity, disability and pain were assessed preoperatively and at 3, 6, 9, 12 and 18 months postoperatively using the Toronto Western Spasmodic Torticollis Rating Scale (TWSTRS). Severity of head tremor and dysphagia were scored using established rating scales. Additionally, psychosocial function was assessed in a representative subsample of patients (n = 12) using several established questionnaires. Of the 62 patients who were assessed, 22 (35.5%) were not offered surgery, most commonly because of widespread dystonia. Of the remaining 40 patients, 37 have so far had surgery, 31 of whom have been followed up for at least 1 year, and 15 for 18 months after surgery (mean follow-up duration 16.7 months). Using the TWSTRS global outcome score, 68% of patients derived functionally relevant improvement at 12 months after surgery. In the entire operated group, total TWSTRS scores were reduced by 30% at 6 and 12 months after surgery (P < 0.0001). The subscores for severity, disability and pain were reduced by 20, 30 and 40%, respectively, at 6 months (P
0.01) and 20, 40 and 30%, respectively, at 12 months (P < 0.01). Pain increased over time, which appeared to result from muscle reinnervation. TWSTRS scores were not significantly improved in the six patients with primary BT treatment failure. Head tremor did not change. There was a significant improvement of body concept, perceived disfigurement, stigma, and quality of life in the 12 patients whose psychosocial function was assessed. Preoperative disability and restriction of head movement were negatively correlated and the initial response to BT treatment positively correlated with global outcome score. Spread or deterioration of dystonia elsewhere in the body occurred in three patients, with unpleasant sensory symptoms in denervated posterior cervical segments occurring in 14. Ten patients developed mild to moderate dysphagia, and two developed severe dysphagia. We conclude that SPD is an effective treatment for patients with secondary, but probably not for those with primary, BT treatment failure. Reinnervation is not infrequent and can compromise outcome. Postoperative morbidity is low, but there is a risk of dysphagia. botulinum toxin treatment failure; cervical dystonia; posterior ramisectomy; selective peripheral denervation
ANOVA = analysis of variance; BT = botulinum toxin; CD = cervical dystonia; SPD = selective peripheral denervation; TWSTRS = Toronto Western Spasmodic Torticollis Rating Scale
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsAppendix IReferences |
|---|
In patients with cervical dystonia (CD), botulinum toxin (BT) injections given locally into active muscles that cause abnormal head postures have been used with great success (Tsui et al. 1986
; Greene et al., 1990; Moore et al., 1991; Anderson et al., 1992; Jankovic and Schwartz, 1995; Kessler et al., 1999).
It is becoming clear, however, that ~5–10% of CD patients do not respond sufficiently to BT injections. This can happen after an initial good response (secondary non-responders) (Jankovic and Schwartz, 1995; Kessler et al., 1999). In this group, resistance to BT is due to BT antibodies in 50–60% of patients (Duane et al., 1995; Jankovic and Schwartz, 1995; Kessler et al., 1999). There are a few CD patients who have no or little response to begin with (primary non-responders), the reasons for which are unclear. Apart from being disabled, CD patients who do not respond to BT treatment are at risk of developing premature degenerative changes in the cervical spine (Chawda et al., 2000), so that alternative treatment options are needed. Even patients who continue to respond well to BT treatment may tire of the necessary visits every 3 months and wish for a more definitive treatment.
Surgery for CD is not a new form of treatment. Before the advent of BT injections a variety of operations had been tried in numerous patients (Keen, 1891; McKenzie, 1924; Dandy, 1930; Putnam et al., 1949; Cooper, 1964; Sorensen and Hamby, 1966; Arseni and Maretis, 1971; Meares, 1971; Laitenen and Vilkki, 1977; Maccabe, 1982; Freckmann et al., 1986; Speelman et al., 1987; Gauthier et al., 1988). The McKenzie operation, developed in the 1920s (McKenzie, 1924), involved bilateral C1 to C3 rhizotomies, but was associated with significant morbidity and mortality (Maccabe, 1982), without evidence of lasting benefit (Meares, 1971) and is therefore no longer recommended. Stereotactic procedures, mainly thalamotomy of the nucleus ventralis oralis internus, have also largely been abandoned (Laitenen and Vilkki, 1977; Maccabe, 1982), although recently some workers have started to investigate the use of pallidal deep brain stimulation (Islekel et al., 1999; Krauss et al., 1999)
The current operation of choice is selective peripheral denervation (SPD), which aims to denervate muscles responsible for abnormal movements while preserving innervation to those that do not participate in the dystonia (Bertrand et al., 1987; Arce and Russo, 1992; Bertrand, 1993; Braun and Richter, 1994; Braun et al., 1995; Krauss et al., 1997; Ford et al., 1998). The technique, originally pioneered by Keen in 1891 (Keen, 1891), was further developed by Bertrand in Montreal who has reported a success rate of 80% in 260 such operations (Bertrand, 1993). Although good surgical outcome with low morbidity has also been reported in other neurosurgical series (Arce and Russo, 1992; Braun and Richter, 1994), little information is available on selection of patients or the roles of severity of dystonia, disability and pain, and other factors in determining the outcome. Two recent retrospective studies addressed these issues (Krauss et al., 1997; Ford et al., 1998), but a prospective study with rigorous assessment of patients has not yet been carried out. Moreover, most surgical series were undertaken before the advent of BT, at a time when surgery was considered one of the first-line treatments for CD due to the lack of effective alternatives. We therefore undertook a prospective study of SPD in patients with primary or secondary BT treatment failure with independent standardized preoperative patient assessment and regular postoperative follow-up.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsAppendix IReferences |
|---|
Subjects
Sixty-two patients with CD and in whom BT injection treatment had failed were assessed between October 1997 and October 2000. The interval between the last BT injection and the assessment was at least 4 months in all patients. Treatment failure was defined as no or only marginal benefit after at least three consecutive sets of BT injections. Eleven patients (18%) never had any significant benefit after BT injections (primary treatment failure), and the remaining 51 had lost their initial favourable response to BT injections after a mean of 6 years (±4.5 SD) (secondary treatment failure). The male : female ratio was 28 : 34 (1 : 1.2), mean age 50 years (±11 SD), mean age of onset 39 years (±12.6 SD) and mean duration of symptoms 11 years (±7 SD). All patients gave informed consent to participation in this study. Procedures were approved by the Joint Medical Ethics Committee of the National Hospital for Neurology and Neurosurgery and the Institute of Neurology.
Peri-operative assessment
The clinical outcome measurements used are given in Table 1.
|
Preoperative assessment
The mean interval between the assessment and surgery was 4.4 months (±3.2 SD).
Clinical examination.
This was carried out by a neurologist specializing in movement disorders (A.M.). A detailed personal, family and drug history was taken in all patients, including trauma before the onset of CD, the presence of a sensory geste and concomitant symptoms, especially tremor and balance or swallowing problems. The initial response to the first set of BT injections was rated retrospectively, based on the patients' accounts and assessment by the treating neurologist as documented in the patients' notes. It was rated according to the global outcome scale of the Toronto Western Spasmodic Torticollis Rating Scale (TWSTRS) (Consky and Lang, 1994) (see Appendix I).
On clinical examination the predominant type of CD (torticollis, laterocollis, retrocollis or anterocollis) was determined, taking into account primarily the degree of restriction of head movements and the primary dystonic head position. For instance, a patient with rotation of the head to the right and tilt to the left was labelled right torticollis when left rotation was more restricted than right tilt and vice versa. The degree of maximum head rotation, head tilt and head flexion or extension was measured using a goniometer. The type of CD was further subdivided into the following groups: (i) torticollis type 1 (torticollis with or without contralateral laterocollis); (ii) torticollis type 2 (torticollis combined with ipsilateral laterocollis); (iii) complex torticollis (combination of torticollis and laterocollis and retrocollis/antecollis of >10° and/or prominent shift of the head); (iv) laterocollis type 1 (laterocollis with or without contralateral torticollis); (v) laterocollis type 2 (laterocollis combined with ipsilateral torticollis) and complex laterocollis (combination of laterocollis and torticollis and retrocollis/antecollis of >10° and/or prominent shift of the head); (vi) simple antecollis; (vii) complex antecollis (combination of antecollis and torticollis and/or laterocollis and/or shift of the head); and (viii) simple retrocollis and complex retrocollis (combination of retrocollis and torticollis and/or laterocollis and/or shift of the head).
Dystonic overactivity of neck muscles was determined by observing and palpating neck muscles during the primary dystonic position, then while the patient was voluntarily moving the head into all directions and then with the patient walking on the spot. Co-contraction of antagonist muscles during voluntary head movements was taken into account.
The severity at the time of the assessment was determined using the TWSTRS, which includes assessment of the dystonic position of the head, neck and shoulder, effectiveness of any sensory geste, how long the patient can keep the head in a straight position and range of head and neck movement (see Appendix I for range of motion subscale indicating the restriction of head movements). The maximum severity score on this scale is 35. Disability and pain were rated according to the TWSTR disability and pain subscales (maximum scores of 30 and 20, respectively). All patients had a baseline video recording following the TWSTRS video protocol.
Severity of head tremor was scored according to the validated clinical rating scale proposed by Bain and colleagues, which defines four main categories of tremor (Bain et al., 1993; see Appendix I). The severity of subjective symptoms of dysphagia were rated according to Comella and colleagues (Comella et al., 1992; see Appendix).
Assessment of psychosocial function.
To assess any changes in psychosocial function associated with surgery, the patients completed a booklet of questionnaires before and after surgery. A covering letter informed the patients that we were interested in their current (during last week or month) personal experiences of living and coping with torticollis. The booklet consisted of the following questionnaires: the Acceptance of Illness Scale (Felton et al., 1984); the Body Concept Scale and Disfigurement Rating (Jahanshahi and Marsden, 1990); the EuroQol Quality of Life Measure (EuroQol Group, 1990; Brooks, with the EuroQol Group, 1996); the Self-esteem Scale (Rosenberg, 1965); the Beck Depression Inventory (Beck et al., 1961); the Beck Anxiety Scale (Beck et al., 1988); the McGill Pain Questionnaire (Melzack, 1975); and the Stigma Scale (MacDonald and Anderson, 1984; Papathanasiou et al., 1997). In addition, the patients rated the present status of their torticollis on a series of 11-point scales by indicating how well they could: (i) voluntarily relax; (ii) voluntarily contract their neck muscles; (iii) voluntarily control their head position; and (iv) move their head in different directions.
EMG.
In the first 25 patients, two-channel needle EMG was used to measure the amplitudes of dystonic and maximal voluntary activities in sternocleidomastoid, splenius capitis and trapezius/semispinalis capitis muscles bilaterally whilst the patient was sitting upright and unsupported and was not trying to oppose the dystonic head movements. The ratio between both amplitudes, the dystonia ratio, was used to quantify dystonic muscle involvement in the examined muscle groups. Where necessary, additional manoeuvres, such as slow active and passive head rotations and application of the sensory geste, were used to identify compensatory muscle activity. Additional muscle pairs were examined if suggested by the clinical picture. Results were displayed as bar charts comparing the dystonia ratios of the muscles studied with the head deviation in the three main head planes. In the remaining 37 patients, polymyography combined with simultaneous videotaping (video-EMG) (Münchau et al., 2001) was carried out. Four neck muscle pairs were recorded. Surface electrodes were used for sternocleidomastoid and trapezius muscles bilaterally and needle electrodes for splenius capitis and levator scapulae. A digital counter, driven by the recording software (resolution 0.1 s), was fixed in view of the video camera allowing time-locking of clinical features seen on video and EMG was recorded on the computer. Video-EMG was used to identify functionally relevant dystonic activity. This was defined as activity during `dystonic spasms' (seen on video), paradoxical activity in antagonist muscles during voluntary head movements causing moderate to severe head restriction (rated on video using the TWSTRS) and muscle activity when symptoms deteriorated during walking.
Additional investigations.
All patients had a CT of the cervical spine from the occipitocervical junction down to the mid/upper cervical spine to detect craniovertebral abnormalities such as os odontoideum and rotatory atlantoaxial dislocation. Also, videofluoroscopic swallow was carried out to test for both symptomatic and also clinically silent swallowing abnormalities.
Planning of surgery.
Eligibility for surgery was determined by consensus between neurologist and neurophysiologist (A.M. and D.D.), neurosurgeon (J.D.P.) and two independent movement disorder specialists (A.J.L., K.P.B.), taking into account the clinical assessment, preoperative EMG and additional investigations.
Patients were excluded from surgery if there were severe balance or severe swallowing problems, risk factors for general anaesthesia, torticollis secondary to structural spine abnormalities, generalized dystonia, predominant antecollis or severely tremulous CD. Once it was concluded that a patient would be eligible for surgery, the surgical procedure was planned according to the pattern of dystonia in neck muscles as determined by clinical examination and EMG.
Postoperative assessment
Patients were formally reassessed by a neurologist (A.M.) 6 weeks, 3, 6, 9 and 12 months after surgery and then every 6 months thereafter. A detailed history regarding change of symptoms or occurrence of new symptoms was taken. The patients were asked in particular about balance problems, sensory symptoms and dysphagia, the last of these being rated in the same way as before surgery. Inquiries were also made as to whether the patients had returned to work and whether they had changed their medication. With regard to the self-assessed outcome after surgery, patients were asked the following question (adopted from Ford et al., 1998): `If you compare the overall benefit after surgery with the best benefit you had following BT injection treatment at any time, which of the two was superior?'
The clinical assessment described above, including videotape of the TWSTRS, was repeated at postoperative assessments. Incomplete denervation and reinnervation were assessed on clinical grounds by inspection and palpation. The latter is reliable as far as the sternocleidomastoid and levator scapulae muscles are concerned, as these muscles lie superficially. It is, of course, more problematic to assess multi-layered posterior cervical muscles and such assessment was restricted to the splenius muscle. Incomplete denervation was defined as lack of atrophy in the target muscle at 6 weeks after surgery and reinnervation as subsequent re-growth of the target muscle that had been atrophic at 6 weeks.
To validate the clinical severity rating of CD, preand 6 months postoperative video segments (of the 32 patients who were followed-up for at least 6 months after surgery) were presented in random order to a blinded reviewer (J.D.O'S), who also rated the severity according to the TWSTRS video protocol. These videosegments were insufficient to enable rating of all components of the TWSTRS in some patients (particularly in relation to the effectiveness of sensory gestes). The global outcome on clinical examination was scored according to the TWSTR outcome scale (see Appendix I).
The booklet of questionnaires assessing psychosocial function, which was completed preoperatively (see above and Table 1
), was again completed an average of 16 weeks (±10.2 SD) and 24 months (±5 SD) following surgery. The patients were also asked to indicate the extent to which they considered their torticollis to be improved with surgery on an 11-point scale (0 = not at all, 100 = greatly improved).
Surgical procedure
The principle of selective peripheral denervation is to denervate dystonic muscles (or muscle groups) while preserving innervation to those that do not participate in dystonia (Bertrand et al., 1987; Arce and Russo, 1992; Bertrand, 1993; Braun and Richter, 1994). There are essentially three components of the operation. A selective denervation of the sternocleidomastoid muscle is achieved by resecting branches of the accessory nerve that leave the main trunk of the nerve from the skull base to the posterior triangle of the neck. A posterior ramisectomy is achieved through a midline posterior incision and sectioning of the rami of C2–C5, 6 or 7 as they emerge around the lateral masses of the vertebrae and of C1, which is identified where it emerges between the vertebral artery and the lateral part of the C1 vertebra. In two patients, attempts were made to denervate levator scapulae muscle by sectioning branches of the dorsoscapular nerve.
Varying combinations of right, left or bilateral selective accessory denervation, posterior ramisectomy and denervation of levator scapulae were selected based upon the clinical and neurophysiological assessments. Surgery was performed under general anaesthesia in a semi-sitting position without muscle relaxation to allow identification of small nerve branches by electrical stimulation during the operation. Myectomy of the upper third of sternocleidomastoid and division of the tendons of levator scapulae were used in adjunct to the denervation.
Immediately after surgery, while still in hospital, patients received intensive daily physiotherapy for 1 week, aimed at improving further the corrected head position. Weekly physiotherapy was continued for 2–3 months. Patients were also instructed to carry out self-exercises on a daily basis.
Statistical analysis
For comparison of two related samples (before and after surgery), paired samples t-test was used for normally distributed interval data. The Friedman's non-parametric repeated measures analysis of variance (ANOVA) was used to compare pre-and postoperative outcome scores. When a significant difference was found in the ANOVA, a pair-wise comparison was performed using Wilcoxon signed rank test with Bonferroni correction. A Wilcoxon signed rank test was also carried out to compare scores rated by the clinical examiner with those of the blinded video reviewer. For ordinal and nominal data comparisons between the patients with primary and secondary BT treatment failure were performed with non-parametric Mann–Whitney U test (and Kruskal–Wallis ANOVA for three or more samples) and 
2 test (Fisher's exact test for small sample sizes), respectively. To measure a possible correlation between different variables and to correlate the TWSTRS severity scores rated by the clinical examiner with those of the blinded video reviewer, Spearman rank-order correlation coefficient was used. For all statistical analyses, an adjusted P value of <0.05 was considered to be significant.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsAppendix IReferences |
|---|
Preoperative assessment and patient selection
Of the 62 patients who were assessed, 22 (35.5%) were not offered surgery. The reasons for patient exclusion are outlined in Table 2
. Thirty-seven have so far undergone surgery, of whom 33 have been followed-up for 3 months, 32 for 6 months, 31 for 12 months, and 15 for 18 months after surgery. Mean follow-up duration was 16.7 months (±6.8 SD). Of the 33 patients who were followed-up for at least 3 months, 22 reported no swallowing problems before surgery, whereas 10 reported mild (score of 1), one moderate (score of 2) and none severe dysphagia after surgery. Among these 33 patients, 27 had idiopathic primary CD, two had axial dystonia and four secondary torticollis, one of whom had tardive dystonia, two had developed symptoms immediately after head (one) or neck (one) trauma and one following an operation for a cervical cyst. Twenty-seven patients had secondary and six primary BT treatment failure. Among the latter six patients were the four with secondary torticollis and one of the two patients with axial dystonia. The initial response following BT injections (determined retrospectively), duration of effective BT treatment and interval between their last BT injection and surgery are shown in Table 3. Preoperative scores are given in Table 4. The only score that differed significantly was head mobility, which was more restricted in patients with primary BT treatment failure. BT antibodies (mouse lethality bioassay) (Hambleton et al., 1992) were determined in 11 patients with secondary BT treatment failure and were found to be positive in five and negative in six.
|
|
|
Seven patients (23.5%) had a positive family history of a movement disorder. Two of these had an uncle who was affected with blepharospasm, two had an aunt or sister with head tremor, one had a cousin with jaw and head tremor, the uncle of one had Parkinson's disease and one patient had a brother with facial tics. Twenty-seven patients (88%) had one or more effective sensory gestes, the most common manoeuvre being touching or approaching the chin or the back of the head with one hand, or pushing the back of the head into a high-backed chair or head rest.
Table 5 shows which surgical procedures were carried out in the different subtypes of CD after dystonic muscles had been identified clinically and with EMG.
|
Outcome after surgery
Global outcome
The global outcome score 3, 6, 12 and 18 months after surgery was 2.8 (±1.2 SD), 2.8 (±1.4 SD), 2.7 (±1.3 SD) and 2.9 (±0.9 SD), respectively [the difference being non-significant (n.s.)]. Patients were divided into two groups, those with an outcome of –1 to +2, i.e. patients who had no or only marginal functional benefit and those with an outcome of +3 to +5, i.e. patients with a functional improvement. At 3, 6 and 12 months after surgery, 73, 63 and 68% of the patients had a functional improvement, respectively (difference n.s.) (Fig. 1
). Ten of the 15 patients (67%) who were followed-up for 18 months had a functional improvement.
|
The global outcome was better and the proportion of patients with functional benefit was significantly larger in the secondary BT treatment failure group compared with those with primary BT failure (Table 4
).
TWSTRS scores
Video rating.
The blinded video reviewer rated severity scores significantly higher than the clinical examiner, both preand postoperatively (Wilcoxon signed rank test, P =0.018 and 0.004, respectively). However, there was no significant difference between raters with regard to the change of the scores after surgery (preoperative scores subtracted from postoperative scores). Moreover, preand postoperative ratings by the examiner and the video reviewer were significantly correlated with each other (Spearman rank-order correlation coefficient, P = 0.002 and P = 0.001 for preand postoperative scores, respectively). In view of this close correlation and incomplete video data on some patients, the TWSTRS ratings made by the clinical examiner were used in subsequent statistical analyses.
TWSTRS scores in patients as a whole group.
Figure 2 shows the total TWSTRS score and Fig. 3 the TWSTR subscale scores before surgery and at 3, 6 and 12 months postoperatively. The total TWSTRS scores at each postoperative assessment were significantly lower than the preoperative score but they did not differ significantly from each other. The mean total TWSTRS score improved by ~30% at both 6 and 12 months after surgery. Severity, disability and pain were all highly significantly reduced at 3, 6 and 12 months after surgery (Fig. 3). There was a trend for pain to increase over time. At 6 months postoperatively, compared with the preoperative scores, there was an approximate 20, 30 and 40% reduction of severity, disability and pain, respectively. At 1 year, severity, disability and pain were reduced by 20, 40 and 30%, respectively. At 18 months, severity, disability and pain in the 15 patients who were followed-up for that time were all significantly reduced by 20, 45 and 30%, respectively (Wilcoxon signed rank test, P = 0.01, P = 0.001, P = 0.02).
|
|
Comparison of TWSTRS scores between patients with primary and those with secondary BT treatment failure.
In the 26 patients with secondary BT treatment failure, the total TWSTRS score was reduced by ~35% at 6 and 12 months (Table 4
) compared with a non-significant improvement of <5% at 6 months (and 15% at 12 months) in the six (five at 12 months) patients with primary treatment failure. Significant improvements were shown in severity (25% at 6 and 12 months), disability (35% at 6 months and 45% at 12 months) and pain (35% at 6 months and 30% at 12 months) in patients with secondary BT treatment failure, but not in patients with primary BT treatment failure. Restriction of head movement scores did not change significantly in either group.
Clinical features
In most patients, the direction of abnormal head posture did not change after surgery, but the degree of abnormal dystonic head rotation and tilt was significantly reduced (mean head rotation: 52° ± 23 SD and 28° ± 20 before and 12 months after surgery, respectively; paired samples t-test, P < 0.0001; mean head tilt: 23° ± 17 and 16° ± 9 before and 12 months after surgery, respectively; P = 0.03). In four patients, the main direction changed (Table 6). On clinical grounds, in none of these patients was the change of head position caused by `new' dystonic overactivity in muscles that were not active before surgery. However, in three patients, in whom the main direction of the CD remained unchanged after surgery, dystonic overactivity became apparent in muscles that had not shown dystonic activity before surgery either clinically or on EMG. In four patients, prominent side shift developed after ipsilateral denervation of the sternocleidomastoid.
|
Changes in psychosocial function
Completed questionnaire booklets were obtained from the last 12 of the patients (10 female, two male) undergoing surgery, who did not differ (all P > 0.05) from the remainder of the sample in terms of age, duration of illness, marital status or employment status. Nor did they differ with regard to preoperative and 3-months postoperative TWSTRS scores, type of CD, head tremor severity or initial response to BT treatment.
At the time of the first completion of the questionnaires after surgery (mean ± SD of 16 ± 10.2 weeks), the patients' average rating of improvement of torticollis was 54%. More specifically, the patients considered their ability to relax voluntarily (paired samples t-test, P = 0.006) and contract (P < 0.015) their neck muscles, control their head position (P < 0.0001) and to move their head in different directions (P = 0.007) to be significantly improved following surgery. Sensory (P = 0.007), affective (P =0.004) and evaluative (P = 0.031) aspects of pain as rated on the McGill Pain Questionnaire were also significantly improved following surgery.
Surgery was associated with significant improvement in a number of measures of psychosocial function. Compared with before surgery, the patients had a significantly less negative body concept (P = 0.042) and perceived themselves as less disfigured (P = 0.012). Self-reported depression (P = 0.017) and perceptions of stigma (P = 0.010) were also significantly lower after surgery (P = 0.017). Scores on the EuroQol Measure of Quality of Life also indicated improvement from before to after surgery (P = 0.020). In contrast, anxiety (P = 0.178), acceptance of illness (P = 0.219) and self-esteem (P = 0.056) were not significantly altered by surgery.
The global outcome score 3 months after surgery had noteworthy correlations with change scores of acceptance of illness (Spearman rank-order correlation coefficient, r =0.63), perceived stigma (r = 0.55) and affective scores on the McGill Pain Questionnaire (r = –0.59). The change in severity of cervical dystonia 3 months after surgery as measured by the total TWSTRS score also had correlations >0.50 with change scores of the total (r = 0.73), sensory (r = 0.72) and evaluative (r = 0.57) measures on the McGill Pain Questionnaire. Because of the relatively small sample size, however, none of these correlations was statistically significant.
At the time of the second completion of the questionnaires after surgery (mean ± SD of 24 ± 5), the change from preoperative levels remained statistically significant for a number of the measures that had also showed improvement 16 weeks after surgery. Patients reported a significant improvement in their ability to voluntarily move their head in different directions (P = 0.013) and control the position of their head (P = 0.006) relative to before surgery. On the McGill Pain Questionnaire, the total (P = 0.013) and the sensory (P = 0.016) aspects of pain were significantly reduced at long-term follow-up relative to preoperative levels.
There were no significant differences between scores obtained an average of 16 weeks and 24 months after surgery for the majority of the measures of psychosocial functioning (P > 0.05), suggesting that the improvements in these measures were maintained at long-term follow-up for body concept, perceived disfigurement, perceived stigma and quality of life. The only exception was that self-reported depression was significantly worse (P = 0.029) at long-term relative to short-term follow-up after surgery.
Other outcome measures
Severity of head tremor did not change significantly after surgery (mean ± SD 2.5 ± 2.4 before and 2.3 ± 2.5 at 12 months after surgery), see Appendix I. However, in two patients head tremor deteriorated noticeably and became the major source of disability. One of the 27 patients with an effective sensory geste reported that his geste (touching the back of his head) was no longer effective postoperatively. The remaining 26 patients stated that their sensory geste was still as effective as before the operation. However, 80% of these patients were using their geste less frequently after surgery.
At 12 months, medication (analgesic and/or anti-dystonic drugs) remained unchanged in 11 out of 31 patients (35%). It was decreased in nine (29%) and increased in three (10%). The other eight patients (25%) were not on any medication before or after surgery.
At 12 months postoperatively, 15 out of the 31 patients who were followed-up for that time had received further BT injection treatment. None of these patients had received BT injections prior to 6 months after surgery. Six of the 15 patients stated that recurrence of pain was the main reason for them to resume BT treatment, eight said that both recurrence of pain and deterioration of head posture were important and one patient said that abnormal head position was the only reason. Eight of these 15 patients reported a moderate response (score of 3 on the TWSTRS global outcome scale), one patient had a mild improvement (score of 2), five patients had a marginal improvement (score of 1) and one patient did not have any benefit (score of 0).
The impact of further BT injection treatment on the outcome 12 months after surgery appears to be limited. The change of mean severity, disability and pain scores and global outcome scores between 6 and 12 months after surgery did not differ significantly between patients who received further BT injection treatment and those who did not (data not shown).
Regarding the self-assessed outcome, at 12 months after surgery, three out of the 31 patients (10%) stated a comparable benefit of surgery and best past response to BT, 21 (67%) estimated that the outcome after surgery was superior and seven (23%) that it was worse. Although their TWSTRS scores did not improve significantly, four out of five patients with primary BT treatment failure (80%) nevertheless said their surgical outcome was superior to their best benefit following BT. One patient stated that the benefit was comparable. For comparison, 16 out of the 26 patients with secondary BT treatment failure (65%) felt their response after surgery was better than their best response to BT injections, seven (27%) said BT treatment was superior and two (8%) that the outcome after surgery and BT was comparable.
Of the 31 patients who were followed-up for 12 months, four were retired. Eleven were not engaged in regular work. None of these patients had gained employment at 12 months after surgery. Sixteen had been working regularly preoperatively. Thirteen of these patients (80%) had resumed their regular employment at 12 months.
Factors determining the outcome after surgery
Gender, age, duration of symptoms, duration of adequate preoperative treatment, head tremor and type of CD were not correlated with global outcome scores. Preoperative disability correlated negatively with global outcome scores (Spearman rank-order correlation coefficient, P = 0.04 and 0.012 at 6 and 12 months, respectively). Range of motion score before surgery correlated negatively with the global outcome score at 3 months (P = 0.002). The initial response to BT treatment score correlated positively with the 3, 6 and 12 months global outcome scores (P = 0.002). It also correlated with the range of motion scores both before and after surgery (P = 0.005 and P < 0.001, respectively). The better the initial response to BT treatment had been, the less restricted head movements were.
Postoperative morbidity, incomplete denervation and reinnervation after surgery
There were no serious life-threatening complications following the operation. One patient with axial dystonia had to stay in hospital for an extended period (39 days) due to a severe wound infection with dehiscence of the scar. A second patient, who had an usual presentation of severe laterocollis following surgery for a neck cyst with erosion of the pinna on the shoulder, had an extended stay (77 days) due to progressive spread of dystonia, which became generalized. Three patients reported transient balance problems (feeling of unsteadiness) and seven reported transient dysaesthesiae (between 2 and 9 months after surgery) in the denervated posterior cervical segments, predominantly C2. No patient reported significant neck muscle weakness compromising head control. Posterior ramisectomy resulted in sensory loss in the corresponding posterior cervical segments in all patients, which persisted to a variable degree. Although most patients were not bothered by this numbness, seven (21%) reported it as discomforting, particularly when lying down. One patient had a transient trapezius paresis that recovered completely. Two patients had worsening of oromandibular dystonia, which persisted at 12 months in one of them.
Seven of the 22 patients without dysphagia preoperatively (in the group of 33 patients who were followed-up for at least 3 months) developed dysphagia after surgery. Four had mild to moderate and one severe transient dysphagia. Two developed mild dysphagia that persisted at 12 months. In five out of 10 patients with pre-existing mild dysphagia before surgery, dysphagia deteriorated. Four developed moderate dysphagia, which was transient in three and persisted at 12 months in the other. One developed transient severe dysphagia. On the other hand, one patient with mild swallowing problems before surgery reported improved swallowing function postoperatively. In the patient with pre-existing moderate dysphagia, swallowing function was unchanged after surgery.
Incomplete denervation was found in two patients, relating to the caudal portion of the sternocleidomastoid in both. Between 6 and 12 months after surgery, clinical evidence of reinnervation was found in 10 out of 22 surgically denervated sternocleidomastoid muscles. In one case it was complete, in five only the caudal cleidal portion had re-grown, and in the remaining four there was some re-growth across the entire length of the muscle. Out of 38 denervated posterior cervical segments, there was re-growth in eight splenii muscles.
Two patients have subsequently had a repeat posterior ramisectomy (6 and 18 months after the initial surgery) because of reinnervation, with a good outcome (score of 4 on the Global Outcome Scale 6 months after re-operation).
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsAppendix IReferences |
|---|
We have carried out the first prospective study of SPD for CD patients with primary or secondary BT treatment failure in which an independent standardized preand postoperative assessment was used.
Overall outcome after selective peripheral denervation
Our main finding was that, using the TWSTRS global outcome score, ~70% of patients had a functionally relevant improvement of their symptoms following surgery (Fig. 1). Also, in the whole series of operated patients there was a significant reduction (30%) of total TWSTRS scores up to 18 months after surgery, which is similar to a previous retrospective study of SPD (Ford et al., 1998). Severity, including the degree of head rotation and head tilt, disability and pain were all significantly improved 6, 12 and 18 months after surgery. Most patients continued to require medication and only 29% could reduce their dosage. Also, 15 patients decided to have further BT injections. In all but one patient, pain was given as a reason for continuing medication and resuming BT injections and abnormal head and neck position was also stated in eight. Continuing BT injection treatment in a subgroup of patients may partly explain why mean severity scores remained stable between 3 and 12 months postoperatively (Fig. 3). However, in view of the relatively modest self-reported benefit of further BT injections and the fact that the change of TWSTRS outcome scores between 6 and 12 months after surgery did not differ significantly between patients who received further BT injection treatment and those who did not, the overall impact of further BT injections on outcome after surgery is likely to be limited.
Further BT injections and continuing medication in the majority of patients did not prevent an increase of pain scores after 3 months postoperatively. Pain scores up to 9 months postoperatively may have been affected by dysaesthesiae present in seven patients after surgery, predominantly in the C2 dermatome. However, as dysaesthesiae had subsided after 9 months in all patients, they would not explain the increased pain scores at 12 months. On the other hand, the increase in pain coincided with the occurrence of reinnervation so that the latter is a likely explanation.
The problem of incomplete denervation and reinnervation
We encountered incomplete denervation in only two patients (caudal portion of the sternocleidomastoid in both cases). However, reinnervation, albeit incomplete in most cases, occurred in ~45% of denervated sternocleidomastoid and 20% of denervated splenii muscles and became apparent 6–12 months after surgery. As deep posterior muscles were not accessible to clinical examination, the frequency and degree of reinnervation might have been underestimated. There are no comparable data from other studies, as regular clinical follow-up with a view to detecting reinnervation has so far not been performed. The fact that in Ford and colleagues' retrospective study (Ford et al., 1998), six of 16 patients required a second or third operation for CD suggests that reinnervation also posed a considerable problem in their group of patients.
The question has to be addressed whether reinnervation in our patients was the result of incomplete sectioning of small nerve branches facilitating later reinnervation. Our use of intraoperative electrical stimulators to identify small branches of the supplying nerves would minimize this possibility; however, the regeneration potential of posterior cervical rami is well known (Braun and Richter, 1994; Braun et al., 1995). Reinnervation of the sternocleidomastoid most often occurred in caudal-cleidal portions of this muscle. It has been suggested that cervical roots contribute to sternocleidomastoid innervation, particularly the cleidal portion (MacKenzie, 1955). Significant reinnervation might have occurred via this route, possibly by direct sprouting of branches from the cervical plexus to the sternocleidomastoid, rather than of branches from the accessory nerve.
In spite of noticeable reinnervation after 6 months in a considerable proportion of our patients, severity scores did not increase between 6 and 12 months. Apparently in most patients reinnervation was incomplete so that overall improvement of head posture was not compromised. Only two patients had recurrence of symptoms to preoperative levels (after an initial postoperative improvement) due to extensive reinnervation in posterior neck muscles; however, repeated surgery resulted in a good outcome in both. On the other hand, as pointed out above, re-growth of dystonic muscles is likely to be responsible for the increase in mean pain scores over time.
It remains to be established whether new surgical techniques can overcome the problem of reinnervation, or whether one simply has to accept that re-growth of nerves will inevitably occur in a proportion of patients who may then need further surgery. The need for re-operation has also been emphasized by others (Krauss et al., 1997; Ford et al. 1998). As the frequency of side effects seems to depend on the extent of surgery (Colbassani and Wood, 1986), staging of surgery may actually be advantageous to avoid complications (Krauss et al., 1997).
Change of muscular activation pattern and head tremor
We observed that, after surgery, the main direction of torticollis changed in four patients. The postoperative change of head posture could be explained on the basis of muscle activation patterns already present preoperatively (Table 6). On the other hand, in three patients dystonic overactivity became apparent in muscles that had not shown dystonic activity before surgery. Similar observations have been made following BT injection treatment (Gelb et al., 1991; Marin et al., 1995; Kanovsky et al., 1997), although the reason for this remains unclear.
As in other series (Krauss et al., 1997), head tremor did not change after surgery, which can be explained by the failure of limited surgical denervation to sufficiently control widespread muscular activity that usually underlies tremulous CD. This is in keeping with responses to BT treatment that are often less satisfactory in patients with tremulous CD (Rivest and Marsden, 1990). Predominantly tremulous CD should be considered a relative contraindication to SPD.
Disability and psychosocial function
In spite of the observed increase in pain after 6 months, improvements in disability scores were maintained and were reduced by 40% at 1 year compared with baseline. With subjective ratings like these, one has to consider a placebo effect, which can be considerable after surgery (Beecher, 1961). The magnitude of improvement in disability that persisted over time suggests that this is unlikely to be the sole reason.
Before undergoing surgery our patients were without adequate treatment for some 2 years. Their baseline total TWSTRS scores were relatively high compared with those of CD patients who continue to respond to BT injections (Brashear et al., 1999). The 25% improvement in severity and 50% improvement in pain (see Fig. 3) 3 months after surgery is therefore likely to have been perceived as a dramatic response with a rapid favourable effect on functional status, as reflected in the global response. This is confirmed by the fact that shortly following surgery the subsample of patients who completed the booklet of questionnaires assessing psychosocial function reported significant improvements in their ability to move and control their head position and reduced pain experience. These patients also perceived themselves as less disfigured, had a more positive attitude towards their body and reported less stigma and depression and an improved quality of life. Once the head deviation and the associated disfigurement and disability are improved with surgery, psychosocial function benefits significantly. Conceivably, in spite of a recurrence of pain after 6 months postoperatively, functional status, once improved, is less likely to undergo short-term changes. Functional status is probably the most important outcome measure as it has recently been shown that only functional status measures reflect therapeutic efficacy for BT treatments (Lindeboom et al., 1998).
The systematic and intense postoperative physiotherapy and continuing self-exercises at home might have contributed to the overall favourable functional outcome after surgery. Whether the functional improvement that seemed to be sustained up to 24 months after surgery will be maintained in the longer-term requires further follow-up. It also remains to be determined whether all the patients who were employed before the operation will eventually return to work (so far 80% have done so) and whether some patients who were unable to work before the operation will be able to start work.
Interestingly, in spite of a lasting significant improvement of most psychosocial aspects at 24 months after surgery, depression scores at that time were significantly higher compared with short-term follow-up. This discrepancy is unclear. One possible explanation is that, although less affected by torticollis postoperatively, patients continue to have a chronic, disabling disease. Patients often had high expectations before surgery, hoping for a more definite treatment. Whereas some of these expectations might have been met immediately after surgery, when there was a very noticeable improvement of symptoms and also intensive postoperative physiotherapy with the prospect of further improvement, 2 years after surgery patients were probably more aware of and had to come to terms with the fact that they continue to suffer from a chronic condition.
Surgery versus BT treatment
Patients' self-assessed outcome after surgery in comparison with their response to BT injections was favourable. Some 70% of patients reported that the overall outcome after surgery was superior to their best response after BT injections and only 23% said it was worse. This is a surprising result, particularly as 74% of patients with secondary BT treatment failure apparently had experienced a good, functionally significant improvement following their first BT injections. Recall bias with an underestimation of the BT response may have played a role. It is well known that the first few BT injections in CD patients who respond to this mode of treatment usually have the most dramatic effect, often leading to a 50% improvement in severity (Kessler et al., 1999). Patients are then generally injected every 3–4 months and tend to deteriorate towards the end of each injection cycle, but usually do not reach their baseline level of severity (Kessler et al., 1999). Consequently, although repeated BT injections continue to improve symptoms, the effect of individual injections becomes less dramatic depending on the injection interval. Moreover, all our patients had eventually failed to respond to BT injections. Therefore, the postoperative improvement might have been more noticeable than the effect of repeated BT injections during the preceding years.
Determinants of surgical outcome
What determined the outcome after surgery? Preoperative disability correlated with global outcome, indicating that patients with high disability scores are less likely to have a good outcome after surgery. The initial response to BT treatment also correlated with post-operative global outcome. Patients with a good initial BT response tended to have a good outcome after surgery. In contrast, symptoms did not significantly improve in patients with primary BT treatment failure, which is in agreement with reports of others (Braun and Richter, 1994; Braun et al., 1995; Ford et al., 1998). However, Krauss and co-workers found no difference in the response between patients with primary and secondary BT treatment failure (Krauss et al. 1997). The reason for this discrepancy is unclear. In our patients with primary BT treatment failure, head movements were significantly more restricted than in those with secondary treatment failure, even before surgery. Severe restriction of head movement was associated with a poorer outcome and was largely unchanged after surgery. The poor outcome in patients with primary BT treatment failure might thus be explained by their relatively fixed head posture, which has been shown to be associated with osteoarthritis of the cervical spine (Chawda et al., 2000). We suggest that severe restriction of head movements should be considered a relative contraindication to surgery.
Postoperative morbidity and effects of surgery on the sensory geste
Postoperative morbidity was low. Posterior ramisectomy resulted in persisting numbness in the corresponding dermatomes (posterior parts), which was perceived as unpleasant and irritating in seven patients. Other side effects included transient unsteadiness (n = 3), transient dysaesthesiae (n = 7), reversible trapezius paralysis (n = 1) and worsening of additional pre-existing oromandibular dystonia (n = 2), which persisted in one case. In one patient, with unusual clinical features before surgery, dystonia spread and became generalized. This phenomenon has been reported previously (Ford et al., 1998), but is poorly understood. Swallowing problems were usually transient and mild, but two patients without (and one patient with mild) preoperative dysphagia before surgery continue to have mild (and moderate) dysphagia 12 months after surgery. Dysphagia may have been caused by damage of anterior rami of cervical roots due to traction during the operation. Conversely, one patient reported that swallowing had improved after surgery, perhaps indicating that dysphagia was caused by an abnormal head position that was corrected by surgery.
Interestingly, only one patient reported loss of effectiveness of his sensory geste (touching the back of the head) after surgery. In all other patients it was still effective. The efficacy of the sensory geste has been interpreted as a result of central modulation of the motor programme by cutaneous inputs due to touch of the face or neck by the patient's arm or hand (Podivinsky, 1968). However, in a number of patients who underwent posterior ramisectomy, their sensory geste (e.g. pushing the head into a high chair) continued to be effective, despite the fact that the sensory supply of posterior cervical segments was interrupted. This indicates that sensory inputs are not a prerequisite for an effective geste manoeuvre. The execution of a motor programme (extending the head or moving the arm) may be equally important (Wissel et al., 1999).
|
Conclusions |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsAppendix IReferences |
|---|
The main conclusions of this study can be summarized as follows. (i) Postoperatively, ~70% of CD patients had a functionally relevant improvement of their symptoms, with severity, disability and pain all significantly improved up to 18 months after surgery. (ii) Pain scores started to increase again 3 months postoperatively, which appears to be largely explained by reinnervation of neck muscles. (iii) Reinnervation, albeit incomplete in most cases, was common and occurred in ~45% of denervated SCM, particularly in the caudal portion and 20% of denervated posterior neck muscles. Repeated surgery is sometimes necessary. (iv) Muscular activation pattern underlying CD can change after surgery. (v) Surgery resulted in a significant and persistent improvement in a number of measures of psychosocial function, including body concept, perceived disfigurement, stigma and quality of life. However, self-reported depression, which was initially improved after surgery, returned to baseline levels at long-term follow-up. (vi) Postoperatively, 80% of patients who were employed before surgery returned to work. (vii) Seventy per cent of patients reported that the overall outcome after surgery was superior to their best response to BT injection treatment in the past. (viii) High preoperative disability, primary BT treatment failure and severe restriction of head movements were associated with a poor outcome after surgery. Head tremor scores did not change. (ix) Overall postoperative morbidity was low, but persisting numbness in posterior cervical segments after posterior ramisectomy can be bothersome and patients are at risk of developing dysphagia. In three patients, there was a spread or deterioration of dystonia elsewhere in the body.
|
Appendix I |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsAppendix IReferences |
|---|
TWSTRS global outcome scale (Consky and Lang, 1994
)–1 = deterioration of symptoms.
0 = no benefit.
1 = minimal or questionable reduction in dystonia and pain with no functional improvement.
2 = mild response with some reduction in dystonia and/or pain, little functional improvement.
3 = moderate response with definite reduction in dystonia and/or pain and some functional improvement.
4 = marked response of dystonia and/or pain, dystonia still evident, excellent functional improvement.
5 = striking improvement with little or no dystonia or pain remaining.
TWSTRS range of motion subscale (Consky and Lang, 1994)
0 = able to move to extreme opposite position.
1 = able to move head well past midline, but not to extreme opposite position.
2 = able to move head barely past midline.
3 = able to move head toward, but not past, midline.
4 = barely able to move head beyond abnormal posture.
Head tremor rating scale (Bain et al., 1993)
–0 = no tremor.
1–3 = noticeable but mild tremor.
4–6 = moderate tremor, which may be bothersome to the patient but does not lead to significant functional impairment.
7–9 = severe tremor present in all head positions.
10 = extremely severe tremor.
Dysphagia rating scale (Comella et al., 1992)
0 = no dysphagia.
1 = mild occasional dysphagia symptoms occurring only with certain foods and without associated choking or coughing.
2 = moderate frequent dysphagia resulting in minor changes in daily diet or occasional choking or coughing.
3 = severe continual dysphagia resulting in major changes in diet, weight loss and/or coughing and choking associated with swallowing.
|
Acknowledgments |
|---|
The authors wish to thank A. Oester-Barkey for her helpful comments. We would also like to thank the reviewers for their detailed and constructive comments. A.M. was supported by the Ernst Jung-Stiftung für Wissenschaft und Forschung in Hamburg, Germany and by the Eugen Brehm Bequest, UK. Technical assistance from Mr M. O'Sullivan during the editing of the videos is gratefully acknowledged.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsAppendix IReferences |
|---|
Anderson TJ, Rivest J, Stell R, Steiger MJ, Cohen H, Thompson PD, et al. Botulinum toxin treatment of spasmodic torticollis. J R Soc Med 1992; 85: 524–9.[Abstract]
Arce C, Russo J. Selective peripheral denervation: a surgical alternative in the treatment for spasmodic torticollis. Review of fifty-five patients [abstract]. Mov Disord 1992; 7 Suppl 1: 128.
Arseni C, Maretsis M. The surgical treatment of spasmodic torticollis. Neurochirurgica (Stuttg) 1971; 14: 177–80.
Bain PG, Findley LJ, Atchison P, Behari M, Vidailhet M, Gresty M, et al. Assessing tremor severity. J Neurol Neurosurg Psychiatry 1993; 56: 868–73.[Abstract]
Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. An inventory for measuring depression. Arch Gen Psychiatry 1961; 4: 561–71.
Beck AT, Epstein N, Brown G, Steer RA. An inventory for measuring clinical anxiety: psychometric properties. J Consult Clin Psychol 1988; 56: 893–7.[ISI][Medline]
Beecher HK. Surgery as placebo: a quantitative study of bias. JAMA 1961; 176: 1102–7.
Bertrand CM. Selective peripheral denervation for spasmodic torticollis: surgical technique, results, and observations in 260 cases. Surg Neurol 1993; 40: 96–103.[ISI][Medline]
Bertrand C, Molina-Negro P, Bouvier G, Gorczyca W. Observations and analysis of results in 131 cases of spasmodic torticollis after selective denervation. Appl Neurophysiol 1987; 50: 319–23.[Medline]
Brashear A, Lew MF, Dykstra DD, Comella CL, Factor SA, Rodnitzky RL, et al. Safety and efficacy of NeuroBloc (botulinum toxin type B) in type A–responsive cervical dystonia. Neurology 1999; 53: 1439–46.
Braun V, Richter HP. Selective peripheral denervation for the treatment of spasmodic torticollis. Neurosurgery 1994; 35: 58–63.[Medline]
Braun V, Richter HP, Schröder JM. Selective peripheral denervation for spasmodic torticollis: is the outcome predictable? J Neurol 1995; 242: 504–7.[Medline]
Brooks R, with the EuroQol Group. EuroQol: the current state of play. Health Policy 1996; 37: 53–72.[ISI][Medline]
Chawda SJ, Münchau A, Johnson D, Bhatia K, Quinn NP, Stevens J, et al. Pattern of premature degenerative changes of the cervical spine in patients with spasmodic torticollis and the impact on the outcome of selective peripheral denervation. J Neurol Neurosurg Psychiatry 2000; 68: 465–71.
Colbassani HJ Jr, Wood JH. Management of spasmodic torticollis. Surg Neurol 1986; 25: 153–8.[Medline]
Comella CL, Tanner CM, DeFoor-Hill L, Smith C. Dysphagia after botulinum toxin injections for spasmodic torticollis: clinical and radiologic findings. Neurology 1992; 42: 1307–10.