1

Instrumented assessment of the effect of Botulinum Toxin-A in the medial 1

hamstrings in children with cerebral palsy 2

3

Bar-On L. PT MSca,b, Aertbeliën E. Ir PhDc, Molenaers G. MD PhDa,d,e, Van 4

Campenhout A. MDa,d,e, Vandendoorent B. PT MScb, Nieuwenhuys A. PT MSca,b, 5

Jaspers E. PT PhDb,f, Hunaerts C. PT MSca; Desloovere K. PhDa, b 6

7

a Clinical Motion Analysis Laboratory, University Hospital Leuven, Pellenberg, 8

Belgium 9

b KU Leuven Department of Rehabilitation Sciences, Leuven, Belgium 10

c KU Leuven Department of Mechanical Engineering, Leuven, Belgium 11

d KU Leuven Department of Development and Regeneration, Leuven, Belgium 12

e Department of Orthopedics, University Hospital Leuven, Pellenberg, Belgium 13

f Neural Control of Movement Lab, ETH Zurich, Switzerland 14

15

Acknowledgements 16

This work was made possible by a grant from the Doctoral Scholarships Committee 17

for International Collaboration with non EER-countries (DBOF) of the Katholieke 18

Universiteit Leuven, Belgium. This work was further supported by a grant from 19

Applied Biomedical Research from the Flemish agency for Innovation by Science and 20

technology (IWT-TBM: grant number 060799); and by an unrestricted educational 21

grant from Allergan, Inc. (USA). 22

23

24

25

26

27

28

2

1. Introduction 29

30

Spasticity is characterized by a velocity-dependent increase in tonic stretch reflex [1] 31

with an accompanying increase in muscle resistance when a muscle is passively 32

stretched [2]. This definition, as well as the methods for spasticity assessment, has 33

been under much debate in the last decade. Nonetheless, neuromuscular tone 34

reduction remains an important treatment modality in children with cerebral palsy 35

(CP) [3]. For example, Botulinum Toxin type-A (BTX-A), injected intramuscularly, 36

causes a temporary reduction in reflex muscle activity by selectively blocking the 37

release of acetylcholine at the cholinergic nerve terminals. Whilst this has been found 38

effective to decrease spasticity in children with CP, there remains a large variability in 39

treatment response [4]. A comprehensive assessment of the effect of BTX-A on 40

spasticity could increase our knowledge of the pathology and improve our 41

understanding of this reported variability. 42

43

In children with CP, the effect of BTX-A is most commonly assessed with clinical 44

scales (Modified Ashworth-MAS [5], or Modified Tardieu Scale-MTS [6]). These 45

scales assess spasticity by subjectively interpreting the resistance felt during passive 46

stretch. Nonetheless, the perceived resistance may be a result of reflex muscle 47

activity as well as of changes in visco-elastic properties of the joint and muscle. The 48

available clinical scales fail to distinguish between both components and are thus not 49

deemed sensitive or valid to quantitatively assess the effect of BTX-A on the stretch 50

reflex. Moreover, they have also been criticized for their low reproducibility and poor 51

accuracy [7,8]. As such, clinical scales have a limited ability to differentiate between 52

3

patients or to explain the response variability after treatment. Instrumented methods 53

could provide a more comprehensive assessment. 54

55

Electromyography (EMG) has been used in adults to quantify the effect of BTX-A on 56

the pathological response during passive muscle stretch [9,10]. Simultaneously 57

assessing muscular resistance using torque sensors provides an integrated (EMG 58

and torque) instrumented measurement method [11]. However, in children with CP, 59

clinically-applicable integrated approaches to assess the effect of BTX-A have only 60

been applied to the upper limb [12], whereas lower limb muscles are most commonly 61

treated. We therefore used an instrumented method, that integrates EMG and torque, 62

as described by Bar-On et al. [13]. The repeatability and discriminate validity to 63

measure spasticity in the medial hamstrings (MEH) in children with CP has previously 64

been shown [13,14]. However, it is yet to be determined whether this instrumented 65

assessment is sensitive to detect treatment efficacy and if it can help understand 66

variability in treatment outcome. 67

68

Therefore, the aim of this study was to quantify and understand the effects of BTX-A 69

injection in treating MEH spasticity in children with CP, using an integrated 70

assessment based on EMG and torque. 71

72

73

2. Method 74

75

Children aged 3-18 years and scheduled for BTX-A in the MEH (Mm. 76

Semitendinosus and Semimembranosus) were recruited from the multidisciplinary 77

4

clinic for patients with CP (University Hospital ***). The exclusion criteria were: 78

presence of ataxia or dystonia; severe muscle weakness (<2+ on the Manual Muscle 79

Test [15]); poor selectivity [6]; bone deformities or contractures hindering neutral 80

alignment; cognitive problems that could impede the measurements; previous lower 81

limb orthopedic surgery (soft tissue or bony procedures); intrathecal Baclofen pump 82

or selective dorsal rhizotomy. Minimal strength production and good selectivity were 83

required because a voluntary contraction was used as an individual reference to 84

evaluate surface EMG (sEMG) signals in previous studies with the same subject 85

group [13,14]. In the current study however, voluntary contractions were expected to 86

be influenced by the BTX-A injections and the normalized sEMG was thus not 87

analyzed. The University Hospitals’ ethical committee approved the experimental 88

protocol and all children’s parents signed an informed consent. 89

90

As part of a regular multilevel BTX-A treatment, muscles to inject and dosages were 91

selected based on standard multidisciplinary evaluation. Injection with BTX-A 92

(Botox®, Allergan Ltd, UK) was done under a short anesthesia and ultrasound was 93

used to confirm needle position. All children underwent casting for a period of 10 94

days (lower-leg with optional removable upper-part used as a knee-extension 95

device), intensive physical rehabilitation as well as orthotic management (day and 96

night) following the BTX-A injections. 97

98

2.1 Data acquisition 99

100

The set-up of the instrumented assessment for the MEH is presented in Figure 1. In 101

children with unilateral CP, only the affected side was tested. In children with bilateral 102

5

involvement, the most involved side was tested. This was defined as the side with the 103

highest MEH MAS-score or, in case of symmetrical MAS-scores, the most severe 104

MTS-score. All assessments were performed prior to injection and 14-70 days after 105

injection, by the same trained assessor. For more details regarding the measurement 106

method, the reader is referred to [13]. 107

108

Four repetitions of passive MEH muscle stretches over the full range of motion 109

(ROM) were carried out at three velocities. Firstly, the knee joint was moved at low 110

velocity (LV) during 5s, followed by a movement at intermediate, medium velocity 111

(MV) during 1s, and finally at high velocity (HV), which was performed as fast as 112

possible. The interval between repetitions was 7s in order to avoid post-activation 113

depression of the electrophysiological response. 114

115

2.2 Data analysis 116

117

A 6th order zero-phase Butterworth bandpass filter ranging from 20-500Hz was 118

applied to filter the raw sEMG signal. The root mean square envelope of the sEMG 119

(RMS-EMG) signal was computed using a low-pass 30Hz 6th order zero-phase 120

Butterworth filter on the squared raw signal. EMG onset, ROM, maximum angular 121

velocity (VMAX), and the net internal joint torque were computed as previously 122

described [13]. 123

124

Repetitions were excluded when passive stretches were performed out of plane, at 125

inconsistent velocities, in case of poor quality sEMG signal (loss of signal, low signal-126

to-noise ratio or obvious artifacts), or when there was indication of antagonist 127

6

activation (rectus femoris sEMG activity). All data analyses were carried out with 128

MATLAB® Software 7.6.0 R2010a. 129

130

2.3 Outcome parameters 131

132

ROM was determined during LV; VMAX during all velocities. All other parameters were 133

calculated at each velocity and were extracted from the RMS-EMG and the computed 134

torque signals. Average RMS-EMG, expressed in mV, was computed as the square 135

root of the area underneath the RMS-EMG time curve, divided by the duration of the 136

time interval considered. The time interval started 200ms prior to the time 137

corresponding to VMAX and ended at the time corresponding to 90% of the full ROM. 138

From the computed torque signal, four instrumented spasticity parameters were 139

developed. Firstly, the amount of work required to stretch the muscle was calculated 140

as the integral of the net internal torque from the joint position at VMAX to 90% of the 141

ROM (referred to as ‘work’ and expressed in J). Torque was additionally analyzed at 142

70° knee flexion, an angle that corresponded to the overall mid-ROM of all children 143

(‘torque’, expressed in Nm). The angle of catch (AOC) was defined as the angle that 144

corresponded to the time of minimum power after maximum power and was 145

expressed as a percentage of the ROM [14]. Finally, the value of the power at the 146

AOC was used to quantify catch severity [14] (‘AOC power’, expressed in W). The 147

AOC and AOC power were calculated from the first HV stretch following the 148

procedure described in [14]. All other parameters were calculated by taking the 149

average of 2-4 repetitions per velocity. To provide a measure of the severity of 150

spasticity, the absolute change between MV and LV (MV-LV) and between HV and 151

7

LV (HV-LV) was also calculated for every parameter (except ROM, AOC, and AOC 152

power). 153

154

155

2.4 Statistical analysis 156

157

All parameters were checked for normal distribution using the Kolmogorov-Smirnov 158

test with p>0.1 indicating a normal distribution. To ensure that the velocity of passive 159

stretches was performed consistently between measurement sessions, VMAX at each 160

velocity was first compared between sessions using a paired samples t-test, or in 161

case of non-normal distributions, a Wilcoxon Matched Pairs Test (WMPT). Next, to 162

evaluate the sensitivity of the parameters to treatment with BTX-A, the average 163

change between pre- and post-treatment sessions was calculated. It was 164

hypothesized that ROM, AOC, and AOC power would increase and that RMS-EMG, 165

torque, and work parameters would decrease post-treatment. Average change 166

between pre- and post-treatment sessions was interpreted in view of the minimal 167

detectable change (MDC). MDC values were calculated from the standard error of 168

measurement (SEM) values reported by Bar-On et al. [13,14] (MDC=SEM*1.645*√2) 169

[16] (Supplementary Material 1). Those parameters whose average change 170

exceeded the MDC were compared between sessions using a paired samples t-test, 171

or a WMPT, as appropriate. 172

173

Finally, to explore the relationships between different outcome parameters, Pearson 174

product-moment correlation coefficients (or Spearman rank correlation coefficients, 175

as appropriate) were computed between all parameters. Correlations <0.20 were 176

8

considered poor; 0.21–0.40 fair; 0.41–0.60 moderate; 0.61–0.80 good; and 0.81–177

1.00, very good [17]. It was hypothesized that the pre-post change in torque 178

parameters would have the highest positive correlations to pre-post change in EMG 179

parameters at HV and at HV-LV. Significance was set at p<0.05. All statistical 180

analyses were performed using Statistica 10 (StatSoft). 181

182

183

3. Results 184

185

Nineteen children with CP participated in the study (Table 1). The mean dose for the 186

MEH was 3.02 U/kg (SD 0.75 U/kg; range: 1-4 U/kg). Assessments were performed 187

on average 9±15 days before and 43±16 days after BTX-A injection (see Figure 2 for 188

a representative example of EMG, torque, position, and power signals pre and post 189

BTX-A). All change pre-post parameters, except RMS-EMG at LV, had a normal 190

distribution. Mean values of all parameters at both sessions can be found in Table 2. 191

Comparison of VMAX between sessions indicated that at HV, muscles were stretched 192

significantly faster during the post-treatment session (increase of 31.5°/sec). ROM 193

increased around 10° post BTX-A, which was significant. At HV stretches, all 194

muscles had an EMG onset during both pre and post BTX-A assessments. RMS-195

EMG parameters significantly decreased post BTX-A at all stretch velocities. Torque 196

and work reduced significantly at HV-LV (decrease of 2.79Nm and 1.18J, 197

respectively), as well as torque at HV (decrease of 3.82Nm). AOC appeared 12.73% 198

further in the ROM and AOC power increased by 3.50W post BTX-A. Both 199

improvements were significant. 200

201

9

The individual and average change pre-post BTX-A for all outcome parameters and 202

their corresponding MDC values can be found in Figure 3.The average change 203

values of ROM at LV (10.90±16.16°), RMS-EMG at HV (0.012±0.011mv), RMS-EMG 204

at HV-LV (0.008±0.009mv), torque at HV-LV (2.79±3.27Nm), and AOC 205

(12.73±16.31%) were larger than their corresponding MDC values (Supplementary 206

Material 1). 207

208

At HV-LV, good correlations were found between RMS-EMG change and torque 209

change pre-post BTX-A (r=0.52), between AOC change and torque change pre-post 210

BTX-A (r=0.58), and a moderate correlation between ROM change and work change 211

pre-post BTX-A (r=0.45). At LV, a moderate correlation was found between ROM 212

change and work change pre-post BTX-A (r=0.45). 213

214

215

4. Discussion 216

217

This study provides a clinically-applicable, instrumented method to quantify the 218

response to BTX-A in spastic MEH in children with CP. Selected parameters, 219

extracted from EMG and torque, were shown to be sensitive to measure effect post 220

BTX-A. 221

222

Clinical spasticity assessments, such as the MAS and MTS, have been criticized for 223

their poor reliability [7,8] and questionable sensitivity in identifying the response to 224

treatment [18]. Moreover, the limited range of the ordinal scoring of the MAS results 225

in patients being clustered into broad severity groups [19]. Although the MTS has a 226

10

smaller gradation, Fosang et al. reported intra-assessor measurement errors up to 9° 227

in the MEH [20]. This results in an MDC value of 21°, which is higher than the 228

average reported change post BTX-A ranging from 2-12° [21,22]. In this study, on the 229

other hand, parameters from the instrumented assessment provided not only 230

sensitive continuous data, but also captured higher variable levels of response to 231

BTX-A treatment. Understanding this variability could enhance treatment delineation 232

and ensure more targeted and individualized anti-spasticity care. 233

234

Clinical tests in isolation cannot discern the relative contributions of neural and non-235

neural components of muscle tone. By integrating electrophysiological and 236

biomechanical parameters, a more comprehensive assessment was achieved. 237

Parameters investigating the change between velocities are able to capture velocity-238

dependent spasticity as defined by Lance [1]. These proved most sensitive to 239

treatment with BTX-A, with an average of 53% reduction in velocity-dependent RMS-240

EMG and a 47% reduction in torque. The moderate correlation between the change 241

in RMS-EMG% and in torque post BTX-A confirms that the decrease in torque is 242

partially influenced by velocity-dependent neurogenic factors. 243

244

All injected muscles showed an increased electrophysiological response to passive 245

HV stretches, indicating that spasticity was correctly diagnosed in all children. In 14 246

of the 19 muscles that were tested, we also found an EMG onset during LV stretches. 247

This may imply the presence of position- or muscle length-dependent spasticity, as 248

also reported in stroke [23] and spinal cord injured [24] patients. It has been 249

suggested that with increasing muscle length, group II afferent neurons activate 250

baseline muscle spindle activity, which in turn lowers the threshold, but not the 251

11

amplitude of the stretch reflex [25]. In the current study, the torque parameters at LV 252

did not change post BTX-A and the decrease in EMG at LV was clinically not relevant 253

(below the MDC). Furthermore, at LV, the change in torque was not correlated to the 254

change in EMG. This suggests that the low intensity muscle activity present during 255

LV stretches does not contribute towards the simultaneously measured torque and 256

that neither parameter is affected by BTX-A treatment. Torque parameters at LV are 257

thus believed to represent intrinsic stiffness (due to secondary changes of the spastic 258

muscle) rather than the neural components of tone [26]. On the other hand, the 259

change pre-post BTX-A in ROM at LV did significantly exceed the MDC. Since little 260

significant effect of BTX-A on intrinsic stiffness has been reported [27], and casting is 261

known to alter the extensibility of muscle and joint, this increased ROM is thought to 262

reflect the effect of post-injection casting. Alhusaini et al. [27] have also reported 263

increased ROM and unchanged intrinsic stiffness due to BTX-A during low-velocity 264

stretches of the gastrocnemius in children with CP. We conclude that it is important to 265

distinguish those patients with increased intrinsic stiffness from those with increased 266

reflex-related torque. This will already help determine the optimal treatment modality 267

for the individual child. 268

269

The AOC and its power value increased post BTX-A, suggesting an increase in the 270

velocity-threshold and a reduction in the catch severity after treatment. However, it 271

should be noted that Wu et al. have warned that the AOC position is positively 272

correlated to the velocity of stretch, with later catches occurring the higher the 273

velocity [28]. Since in the current study, the MEH muscles were stretched at a 274

significantly faster velocity post BTX-A, this could have affected the results. We 275

believe that considering the intensity with which the catch occurs, together with the 276

12

position, improves the interpretation of the spastic catch. However the repeatability of 277

this parameter should be enhanced. 278

279

In the current study, 12 children showed an improvement in RMS-EMG at HV-LV that 280

exceeded MDC, 6 were within the positive and negative MDC values, and 1 showed 281

increased RMS-EMG activity (i.e. worsening of spasticity) that was smaller than the 282

negative MDC value (Figure 3). In comparison, Pandyan et al. [18] measured RMS-283

EMG during fast passive stretches of the biceps muscle in stroke patients before and 284

after BTX-A injections. In accordance to our findings, they reported large response 285

variability, whereby 9 of their 14 subjects had decreased and 4 had increased RMS-286

EMG post BTX-A. They did not however quantify the MDC of the RMS-EMG 287

parameter, which may have led to an overestimation of responders and non-288

responders. In fact, to the best of our knowledge, no study has used information on 289

the measurement error of instrumented spasticity assessments to interpret the effect 290

of BTX-A, making comparisons difficult. 291

292

Despite the sensitivity of the instrumented measurement method, some 293

methodological limitations need to be considered. A first limitation is the significantly 294

higher stretch velocity during post-treatment sessions. However, despite faster 295

stretching post BTX-A, there was still an overall reduction in the spasticity 296

parameters. Furthermore, as has been reported by Chen et al. [12], BTX-A can 297

increase the velocity threshold of the spastic muscle, which could have accounted for 298

the increased VMAX post BTX-A. The VMAX post BTX-A was also closer to the velocity 299

at which the MEH of typically developing children is stretched [13]. A second possible 300

limitation is the lack of EMG normalization. This may have accounted for some of the 301

13

response variability seen among children. However, since BTX-A is known to affect 302

strength, normalization to e.g. maximum voluntary contraction is not suitable and will 303

only increase variability [29]. To ensure reliability of the calculated parameter and 304

minimize variability, a thoroughly standardized electrode application was applied. 305

This standardized procedure without normalization resulted in reliable RMS-EMG 306

parameters (Supplementary Material 1). Finally, while this study investigated the 307

effect of BTX-A on passive spasticity, it is also recommended to capture the effect of 308

BTX-A on functional activities (e.g. walking). More specifically, as only weak to 309

moderate correlations between clinical spasticity scores and gait parameters have 310

been reported [30], it would be useful to explore correlations between parameters 311

from instrumented tests and gait analysis. 312

313

In conclusion, the current study proposes an instrumented method to quantify the 314

effect of BTX-A on MEH spasticity in children with CP. Spasticity parameters that 315

were sensitive to treatment and larger than the MDC were identified. These could 316

potentially be used to categorize subjects according to the level of response and thus 317

assist in treatment planning. This consolidates the clinical validity of the proposed 318

method and opens up possibilities of exploring the effects in other muscles or of other 319

tone-reducing treatments such as selective dorsal rhizotomy and intrathecal baclofen. 320

Furthermore, combining multiple, integrated parameters was found superior over 321

interpreting a single parameter obtained from an isolated signal to assess treatment 322

efficacy. Increased torque at LV, representing intrinsic rather than reflex-related 323

stiffness, did not change post BTX-A. However, the large response variability among 324

children requires further studies using objective, instrumented measurements that 325

assess the effect of BTX-A. A better understanding of factors that determine 326

14

treatment outcome will allow individualized treatment planning and increase the 327

functional potential of children with CP. 328

329

330

Conflict of interest 331

There were no conflicts of interest. 332

333

334

Acknowledgements 335

This work was made possible by a grant from the Doctoral Scholarships Committee 336

for International Collaboration with non EER-countries (DBOF) of the Katholieke 337

Universiteit Leuven, Belgium. This work was further supported by a grant from 338

Applied Biomedical Research from the Flemish agency for Innovation by Science and 339

technology (IWT-TBM: grant number 060799); and by an unrestricted educational 340

grant from Allergan, Inc. (USA). 341

342

343

References 344

345

[1] Lance JW. The control of muscle tone, reflexes, and movement: Robert 346

Wartenbeg Lecture. Neurology 1980;30(12):1303. 347

[2] Sanger TD, Delgado MR, Gaebler-Spira D, Hallett M, Mink JW. Classification and 348

Definition of Disorders Causing Hypertonia in Childhood. Pediatrics 2003;111:89–97. 349

15

[3] Graham HK, Aoki KR, Autti-Rämö I, Boyd RN, Delgado MR, Gaebler-Spira DJ, et 350

al. Recommendations for the use of botulinum toxin type A in the management of 351

cerebral palsy. Gait and Posture 2000;11:67–79. 352

[4] Desloovere K, Schörkhuber V, Fagard K, Van Campenhout A, De Cat J, Pauwels 353

P, et al. Botulinum toxin type A treatment in children with cerebral palsy: evaluation of 354

treatment success or failure by means of goal attainment scaling. European Journal 355

of Paediatric Neurology 2012;16:229–36. 356

[5] Bohannon RW, Smith MB. Interrater reliability of a modified Ashworth scale of 357

muscle spasticity. Physical Therapy 1987;67:206–7. 358

[6] Boyd RN, Graham HK. Objective measurement of clinical findings in the use of 359

botulinum toxin type A for the management of children with cerebral palsy. European 360

Journal of Neurology 1999;6:23–35. 361

[7] Haugh a B, Pandyan a D, Johnson GR. A systematic review of the Tardieu Scale 362

for the measurement of spasticity. Disability and Rehabilitation 2006;28:899–907. 363

[8] Platz T, Eickhof C, Nuyens G, Vuadens P. Clinical scales for the assessment of 364

spasticity, associated phenomena, and function: a systematic review of the literature. 365

Disability and Rehabilitation 2005;27:7–18. 366

[9] Cousins E, Ward AB, Roffe C, Rimington LD, Pandyan AD. Quantitative 367

measurement of poststroke spasticity and response to treatment with botulinum toxin: 368

a 2-patient case report. Physical Therapy 2009;89:688–97. 369

[10] Marinelli L, Trompetto C, Mori L, Vigo G, Traverso E, Colombano F, et al. 370

Manual linear movements to assess spasticity in a clinical setting. PloS One 371

2013;8:53627. 372

[11] Lee H-M, Chen J-JJ, Wu Y-N, Wang Y-L, Huang S-C, Piotrkiewicz M. Time 373

course analysis of the effects of botulinum toxin type a on elbow spasticity based on 374

16

biomechanic and electromyographic parameters. Archives of Physical Medicine and 375

Rehabilitation 2008;89:692–9. 376

[12] Chen J-JJ, Wu Y-N, Huang S-C, Lee H-M, Wang Y-L. The use of a portable 377

muscle tone measurement device to measure the effects of botulinum toxin type a on 378

elbow flexor spasticity. Archives of Physical Medicine and Rehabilitation 379

2005;86:1655–60. 380

[13] Bar-On L, Aertbelien E, Wambacq H, Severijns D, Lambrecht K, Dan B, et al. 381

Clinical measurement to quantify spasticity in children with cerebral palsy by 382

integration of multidimensional signals. Gait and Posture 2013;38:141-7. 383

[14] Bar-On L, Aertbeliën E, Molenaers G, Bruyninckx H, Monari D, Jaspers E, et al. 384

Comprehensive quantification of the spastic catch in children with cerebral palsy. 385

Research in Developmental Disabilities 2012;34:386–96. 386

[15] Daniels L, Worthingham C. Muscle testing techniques of manual examination. In 387

Muscle testing techniques of manual examination. 4th ed. Philadelphia: WB 388

Saunders; 1986:4–26. 389

[16] De Vet HC, Terwee CB, Ostelo RW, Beckerman H, Knol DL, Bouter LM. Minimal 390

changes in health status questionnaires: distinction between minimally detectable 391

change and minimally important change. Health and Quality of Life Outcomes 392

2006;4:54. 393

[17] Katz JN, Larson MG, Phillips CB, Fossel AH, Liang MH. Comparative 394

Measurement Sensitivity of Short and Longer Health Status Instruments. Medical 395

Care 2013;30:917–25. 396

[18] Pandyan AD, Vuadens P, Van Wijck FM, Stark S, Johnson GR, Barnes MP. Are 397

we underestimating the clinical efficacy of botulinum toxin (type A)? Quantifying 398

17

changes in spasticity, strength and upper limb function after injections of Botox to the 399

elbow flexors in a unilateral stroke population. Clinical Rehabilitation 2002;16:654–400

60. 401

[19] Condliffe EG, Clark DJ, Patten C. Reliability of elbow stretch reflex assessment 402

in chronic post-stroke hemiparesis. Clinical neurophysiology 2005;116:1870–8. 403

[20] Fosang AL, Galea MP, McCoy AT, Reddihough DS, Story I. Measures of muscle 404

and joint performance in the lower limb of children with cerebral palsy. 405

Developmental Medicine and Child Neurology 2003;45:664–70. 406

[21] Kelly B, MacKay-Lyons MJ, Berryman S, Hyndman J, Wood E. Assessment 407

protocol for serial casting after botulinum toxin a injections to treat equinus gait. 408

Pediatric Physical Therapy 2008;20:233–41. 409

[22] Scholtes VA, Dallmeijer AJ, Knol DL,Speth LA, Maathuis CG, Jongerius PH, et 410

al. Effect of multilevel botulinum toxin a and comprehensive rehabilitation on gait in 411

cerebral palsy. Pediatric Neurology 2007;36:30–9. 412

[23] Malhotra S, Cousins E, Ward A,Day C, Jones P, Roffe C, et al. An investigation 413

into the agreement between clinical, biomechanical and neurophysiological 414

measures of spasticity. Clinical Rehabilitation 2008;22:1105–15. 415

[24] Van der Salm A, Veltink PH, Hermens HJ, Ijzerman MJ, Nene A V. Development 416

of a new method for objective assessment of spasticity using full range passive 417

movements. Archives of Physical Medicine and Rehabilitation 2005;86:1991–7. 418

[25] Burke D, Gillies JD, Lance JW. The quadriceps stretch reflex in human 419

spasticity. Journal of Neurology, Neurosurgery, and Psychiatry 1970;33:216–23. 420

[26] Foran JRH. Review Structural and mechanical alterations in spastic skeletal 421

muscle. Developmental Medicine and Child Neurology 2005:713–7. 422

18

[27] Alhusaini A a a, Crosbie J, Shepherd RB, Dean CM, Scheinberg A. No change in 423

calf muscle passive stiffness after botulinum toxin injection in children with cerebral 424

palsy. Developmental Medicine and Child Neurology 2011;53:553–8. 425

[28] Wu Y-N, Ren Y, Goldsmith A, Gaebler D, Liu SQ, Zhang LQ. Characterization of 426

spasticity in cerebral palsy: dependence of catch angle on velocity. Developmental 427

Medicine and Child Neurology 2010;52:563–9. 428

[29] Phadke CP, Ismail F, Boulias C. Assessing the neurophysiological effects of 429

botulinum toxin treatment for adults with focal limb spasticity: a systematic review. 430

Disability and Rehabilitation 2012;34:91–100. 431

[30] Desloovere K, Molenaers G, Feys H, Huenaerts C, Callewaert B, Van de Walle 432

P. Do dynamic and static clinical measurements correlate with gait analysis 433

parameters in children with cerebral palsy? Gait and Posture 2006;24:302–13. 434

435

TABLES 436

Table 1. Children’s characteristics 437

Children’s characteristics (n=19)

Mean age (SD) (years) 7.20 (3.09)

Male/female (n) male: 10, female: 9

Diagnosis (n)

- Unilateral involvement

- Bilateral involvement

3 right hemiplegia, 3 left hemiplegia

11 diplegia, 2 quadriplegia

GMFCS (I-IV) (n) I: 6, II: 9, III: 3, IV: 1

MAS MEH (0-4) (number of muscles) 0: 1, 1: 1, 1+: 4, 2: 11, 3: 2

Average MTS MEH (SD) -70.47˚ (26.84˚)

19

GMFCS: Gross Motor Function Classification Score; MAS MEH: Modified Ashworth 438

Score of the medial hamstrings from the pre-treatment session; MTS MEH: average 439

Modified Tardieu Score of the medial hamstrings from the pre-treatment session. 440

441

Table 2. Average and standard deviation values of spasticity parameters pre and 442

post BTX-A treatment. 443

Pre BTX-A Post BTX-A p

ROM (°) LV 73.77 (11.02) 84.67 (13.38) 0.009

VMAX (°/sec) LV 31.21 (10.05) 33.56 (11.26) 0.495

MV 82.83 (32.42) 76.73 (23.85) 0.486

HV 281.43 (39.98) 312.96 (43.88) 0.007

RMS-EMG (µv) LV a4.44 (7.31) a3.12 (7.60) a0.040

MV 14.30 (12.20) 8.67 (5.78) 0.025

HV 25.10 (12.01) 13.40 (8.62) <0.001

MV-LV 6.20 (5.75) 2.93 (3.68) 0.001

HV-LV 16.27 (9.94) 8.31 (7.16) <0.001

Torque (Nm) LV 3.19 (2.47) 2.57 (3.65) 0.535

MV 3.92 (3.44) 2.89 (3.00) 0.316

HV 10.49 (5.66) 6.67 (3.54) <0.001

MV-LV 0.73 (1.67) 0.26 (0.93) 0.244

HV-LV 7.10 (4.14) 4.31 (2.12) 0.002

Work (J) LV 2.85 (2.93) 2.44 (2.06) 0.574

MV 3.35 (2.41) 3.68 (3.37) 0.674

HV 6.57 (3.39) 6.20 (5.01) 0.692

MV-LV 1.11 (1.34) 0.65 (0.60) 0.078

20

HV-LV 4.32 (2.72) 3.14 (2.55) 0.020

AOC (%) HV 79.77 (12.62) 92.50 (9.30) 0.003

AOC power (W) HV -5.94 (6.51) -2.44 (4.34) 0.007

LV: low velocity stretch; MV: medium velocity stretch; HV: high velocity stretch; MV-444

LV: medium velocity stretch minus low velocity stretch; HV-LV: high velocity stretch 445

minus low velocity stretch; ROM: range of motion; VMAX: maximum angular velocity; 446

Torque: torque at 70˚ knee flexion; AOC: angle of catch defined as the angle 447

corresponding to the time of the first minimum power value after the time of the 448

maximum power, expressed as a percentage of the full range of motion; AOC power: 449

the power value at AOC. a Due to the skewed distribution of this parameter, median 450

and inter quartile ranges are provided and compared using the Wilcoxon Matched 451

Pairs Test. p<0.05 452

1

FIGURES 453

454

455

Figure 1. Instrumented spasticity assessment of the medial hamstrings muscle: test 456

starting position, direction of stretch (white arrow) and instrumentation. (1) a six DoF 457

force-sensor attached to a shank orthosis on the posterior aspect of the lower leg 458

(torque measurement); (2) two inertial measurement units (joint angle measurement); 459

and (3) surface electromyography (sEMG) of the medial hamstrings and rectus 460

femoris (muscle activity measurement). sEMG data from the rectus femoris were 461

utilized to ensure no active assistance of the patient during the passive stretches. 462

3

2

1

1

463

Figure 2. Pre (top row) and post (bottom row) BTX-A measurement of a child with spastic CP: RMS EMG-time (a, e), torque-464

position (b, f), position-time (c, g), and power-time graphs (d, h) during low (pink/ light gray), medium (green/gray), and high 465

(blue/black) velocity stretches. The position of the angle of catch (AOC) is indicated on the position- and power-time graphs. 466

467

1

468 469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

486

487

-10

-5

0

5

10

15

20

25

30

RM

S-E

MG

HV

-LV

(µv

)

-30

-20

-10

0

10

20

30

40

50R

OM

LV

(de

g)

-30-20-10

01020304050

AO

C H

V (

%)

-15

-10

-5

0

5

10

15

20

Pow

er H

V (

W)

(j)

(c)

(i)

-6

-4

-2

0

2

4

6

Wor

k H

V (

J)

-6

-4

-2

0

2

4

6

8

Wor

k H

V-L

V (

J)

(h)

(a) -100-80-60-40-20

020406080

100

VM

AX

HV

(de

g/se

c)

(b)

(d)

-6-4-202468

1012

Tor

que

HV

(N

m)

-4

-2

0

2

4

6

8

10

12

Tor

que

HV

-LV

(N

m)

(f) (e)

(g)

-15-10-505

101520253035

RM

S-E

MG

(µv)

2

Figure 3. Change between pre and post BTX-A measurements in: a) range of motion 488

(ROM) at low velocity (LV); b) maximum angular velocity (VMAX) at high velocity (HV); 489

c) root mean square electromyography (RMS-EMG) at HV; d) change in RMS-EMG 490

between HV and LV (HV-LV); e) torque at 70˚ knee flexion at HV; f) torque at 70˚ 491

knee flexion at HV-LV; g) work at HV; h) work at HV-LV; i) relative position of the 492

angle of catch (AOC) at HV; and j) the power value at the AOC at HV. Each diamond 493

represents the individual value of a MEH (per child). The bar represents the mean 494

value for all muscles. The dashed horizontal lines represent the minimal detectable 495

change values. 496