Eur Spine J DOI 10.1007/s00586-014-3267-1
ORIGINAL ARTICLE
Three-dimensional gait analysis outcomes at 1 year following decompressive surgery for cervical spondylotic myelopathy Ailish Malone • Dara Meldrum • Ciaran Bolger
Received: 6 December 2013 / Revised: 25 February 2014 / Accepted: 25 February 2014 Ó Springer-Verlag Berlin Heidelberg 2014
Abstract Purpose Gait impairment is an important feature of cervical sponydylotic myelopathy (CSM) as it can have a detrimental effect on function and quality of life. The aim of this study was to measure changes in gait in people with CSM following surgical decompression. Methods Thirteen participants with clinical and radiological evidence of CSM underwent three-dimensional gait analysis, using a full lower limb kinematic, kinetic and electromyography protocol, before and 12 months after decompressive surgery. Results No significant post-operative changes were detected in temporal–spatial or kinematic parameters. Kinetic data showed significant improvements in knee power absorption [mean improvement, 0.42 watts per kilogram (W/kg)], ankle plantarflexor moment (0.1 Nm/ kg) and ankle power generation (0.55 W/kg). Electromyography showed a 4.7 % increase in tibialis anterior activation time. Conclusions These findings indicate that improvement in locomotor function can be achieved after surgery. Future
A. Malone (&) Gait Analysis Laboratory, Central Remedial Clinic, Clontarf, Dublin 3, Ireland e-mail: [email protected] A. Malone D. Meldrum School of Physiotherapy, Royal College of Surgeons in Ireland, 123 St. Stephen’s Green, Dublin 2, Ireland e-mail: [email protected] C. Bolger Department of Neurosurgery, Beaumont Hospital, Dublin 9, Ireland e-mail: [email protected]
studies should explore the potential for further recovery of gait through targeted neuro-rehabilitation. Keywords Cervical myelopathy Gait analysis Electromyography Biomechanics Surgery
Introduction Surgery is widely used as a treatment for cervical spondylotic myelopathy (CSM), particularly in people whose symptoms show neurological progression over time. The aim of surgical intervention is to stabilise the disease by minimising spinal cord compression at the spondylotic levels, thereby preventing further neurological deterioration [1]. A recent prospective study found significant improvements in function, disability and quality of life 1 year after decompressive surgery for CSM [2]. However, surgery itself carries numerous risks including dysphagia, instrumentation failure or malposition, infection, and damage to neural structures, with an overall prevalence of adverse events of 18.7 % [2]. For many people with CSM, the decision to undergo surgery can be a difficult one and must be weighed against the possibility of further irreversible neurological deterioration if the cord is not decompressed [3]. Gait impairment is one of the cardinal symptoms of CSM. Our previous work determined the key kinematic, kinetic and electromyographic features that characterise gait in people with CSM compared to age- and gendermatched healthy controls [4, 5]. There is some evidence from follow-up studies that gait speed can improve after surgery by up to 0.15 m/s [6, 7]. A more detailed examination of gait before and after surgical intervention is required to assess the true potential for recovery, as well as
123
Eur Spine J
Methods
to-noise ratio of at least 50 dB, amplified with a gain range of 2,000–13,200 and sampled into a PC at 1,000 Hz using a 32-channel DI-720 analogue to digital convertor with 12-bit resolution (DATAQ Instruments Inc., Akron, Ohio, USA). The assessment consisted of a static trial for calibration and a warm-up trial, followed by a number of barefoot walking trials at self-selected speed along a 12-m overground walkway. The assessment continued until ten trials with good-quality data, including five left and right force plate strikes, had been achieved. Rest breaks were provided as required to avoid fatigue.
Recruitment
Data processing
People with CSM were recruited from a neurosurgical spinal assessment clinic at a national neurosurgical hospital over a 2-year period from December 2008 to December 2010. The following inclusion criteria were applied: (1) aged 18 years or over; (2) able to give informed consent; (3) able to mobilise to at least 10 m without assistance of another person (but using aids if necessary); (4) clinical and radiological evidence of CSM. Patients were excluded if they were affected by any of the following: (1) severe respiratory or cardiac disease hindering safe mobilisation; (2) history of neurological disorders with persistent deficit; (3) symptomatic musculoskeletal problems affecting gait; (4) tandem lumbar spine stenosis; (5) previous surgical decompression for CSM. The severity of myelopathy was assessed using the Nurick scale and modified Japanese Orthopaedic Association (mJOA) score.
Gait data were processed using VICON’s Plug-in GaitÒ software. The average of ten trials was used to represent the gait pattern of each participant pre- and post-surgery. Kinematic data were derived for pelvis, hip, knee and ankle in three planes. Kinetic data for joint moments and powers were derived for the hip in the sagittal and coronal planes, and for the knee and ankle in the sagittal plane. EMG signals from each gait trial were imported into MATLAB (The Mathworks Inc., Natick, MA, USA) and filtered with a fourth-order Butterworth low-pass filter of 400 Hz and a second-order Butterworth high-pass filter of 25 Hz, applied in forward and reverse directions to remove motion artefact [10]. Muscle activity was detected using an algorithm, previously validated in this population, which identifies changes in amplitude and frequency from its baseline non-active state to determine contraction [11]. Duration of muscle activity and co-activation between agonist and antagonist muscles was reported as a percentage of gait cycle duration. The EMG amplitude of each burst of muscle activity was extracted by calculating the root-mean square (RMS) of the signal over a 30-ms window with 20-ms overlap [12]. Signals were normalised by expressing the RMS amplitude at each time point as a percentage of the maximum amplitude measured for that muscle during gait [13, 14].
the clinical significance of that recovery. This could provide further information for surgeons and patients on the expected outcomes of surgery and contribute to the expanding base of knowledge on the potential for changes in gait following spinal cord injury. The aim of this study was to determine changes in temporal–spatial, kinematic, kinetic and electromyography (EMG) variables in gait in people with CSM 1 year following surgery.
Gait analysis protocol Three-dimensional gait analysis (3DGA) was conducted prior to and 12 months after surgery using a VICONÒ250 five-camera Motion Analysis System (VICON, Oxford, UK) and an integrated Kistler multi-component force plate (Kistler, Winterthur, Switzerland). Kinematic and kinetic data were collected according to a standard protocol involving the placement of 15 reflective markers to anatomical landmarks on the lower limbs, tracked at 50 Hertz (Hz) by the VICON system [8]. Surface EMG signals were recorded during gait from rectus femoris (RF), biceps femoris (BF), tibialis anterior (TA) and medial head of gastrocnemius (MG), using double-differentiated pre-amplified stainless steel electrodes with a common mode rejection ratio of 100 decibels (dB) at 65 Hz. The skin underlying each electrode was shaved and cleaned to reduce impedance. Electrode placement followed the guidelines of SENIAM [9]. A reference electrode was placed over C7. Signals were collected across a bandwidth of 20–500 Hz with a signal-
123
Statistical analysis Reports have considered the minimal clinically important difference in gait speed to be 0.1 m/s as this shows a good correlation with the ability to carry out functional activities of daily living [15]. A power calculation based on a paired t test found that a sample size of 13 participants would have 90 % power at the 5 % significant level to detect a change in gait speed of 0.1 m/s from pre- to post-operative assessment, with a standard deviation of 0.11 m/s in line with previously published data [6]. Differences in gait parameters from pre- to post-surgery were tested using
Eur Spine J Table 1 Changes in temporal–spatial parameters following surgery Variable
Pre-op Mean
Gait speed (m/s) Cadence (steps/min) Stride length (m)
Post-op SD
Mean
Difference SD
Mean
95 % confidence intervals
p value
SD
1.05
0.32
1.08
0.37
0.03
0.20
-0.09
0.15
0.62
110.30
12.44
111.37
14.87
1.06
17.96
-9.79
11.9
0.83
1.12
0.26
1.14
0.31
0.02
0.11
-0.05
0.09
0.57
Double support duration (% GC)
27.99
6.90
28.72
8.93
0.72
4.88
*
*
Single support duration (% GC)
35.89
2.82
35.68
4.46
-0.22
3.27
-2.19
1.76
0.82
Opposite foot off (% GC)
13.38
2.78
13.79
3.88
0.41
2.83
-1.30
2.12
0.61
Opposite foot contact (% GC) Foot off (% GC)
49.27 63.87
1.31 4.57
49.48 64.39
1.38 4.66
0.21 0.52
1.25 2.20
-0.55 -0.81
0.96 1.85
0.56 0.41
0.16
0.04
0.17
0.05
0.01
0.02
-0.01
0.02
0.26
Step width (m)
*0.75
SD standard deviation, m metres, s seconds, min minute, GC gait cycle * Denotes non-normally distributed variable tested with Wilcoxon signed-rank test
paired t tests where data satisfied the assumption of normal distribution, and using Wilcoxon signed-rank tests for nonnormally distributed data. Analysis was conducted in Stata IC 11 (StataCorp, Texas, USA).
Results Participants Forty-four people met the inclusion criteria over the 24-month recruitment period. Sixteen were excluded due to immobility (four), previous surgery for CSM (four) and pre-existing neurological or orthopaedic conditions (eight). A further ten declined to participate and five did not proceed to surgery. The remaining 13 people with CSM underwent decompressive surgery and gave informed consent to participate. Ten participants were followed up at 12 months as per the study protocol and three participants attended a 6-month follow-up due to time limitations. The mean time from pre-operative assessment to surgery was 2.2 months (range 0–4.9 months). The mean age of the cohort was 56.6 years (range 34–77 years). Myelopathic symptoms were reported for a mean of 59 months prior to surgery (range 9–420 months). The median pre-operative Nurick score was 3 (range 1–4) and median mJOA score was 10 (range 8–13). The most commonly involved cervical levels were C3/ 4, C4/5 and C5/6. Seven participants had an anterior surgical approach, involving an anterior cervical discectomy and fusion (ACDF) in six and an anterior cervical corpectomy and fusion (ACCF) in one case. The remaining six participants had a posterior approach, three of which involved laminectomy and lateral mass plating (LLMP), one, a decompression and lateral mass screws, and two, an occipital–cervical fusion.
Two participants (15 %) suffered post-operative complications. One had a wound infection at the surgical site and subsequent sepsis, and the other experienced instrumentation failure. In both cases, the instrumentation was removed and the cervical spine stabilised in a halo ring and brace. The second participant then underwent a repeat fusion following the removal of the halo brace. These complication rates were in line with a recent large case series [2]. Changes in temporal–spatial parameters Gait speed increased from a mean of 1.05 m/s pre-operatively to 1.08 m/s at the 12-month post-operative assessment. This was not statistically significant (p = 0.62). Similarly, none of the measured temporal–spatial parameters changed significantly (Table 1). Changes in kinematics Kinematic data are shown in Fig. 1. Graphs indicated a post-operative decrease in anterior pelvic tilt, an increase in peak hip extension and an increase in peak knee flexion in swing. These changes were not statistically significant (Table 2). Changes in kinetics Joint moments and powers pre- and post-surgery are shown in Fig. 2. Peak ankle plantarflexor moment increased significantly from 1.34 Nm/kg pre-surgery to 1.44 Nm/kg post-surgery (p = 0.05). There was a significant increase in ankle power at pre-swing from 2.63 to 3.18 W/kg (p = 0.03) and knee power absorption at terminal swing, from 0.66 to 1.08 W/kg (p = 0.01). Knee power in initial swing showed a non-significant increase
123
Eur Spine J Fig. 1 Kinematic data for pelvis, hip, knee and ankle before and after surgery. HC healthy control data
from 0.69 to 0.93 W/kg (p = 0.099). Similarly, hip power at pre-swing increased from 0.89 to 1.24 W/kg (p = 0.098) (Table 3). All changes exceeded previously calculated standard errors of measurement for this population [16].
significant change in timing (Table 4). Similarly, there were no changes in the amplitude of muscle activity (Table 5).
Changes in electromyography
The median post-operative Nurick score was 2 (range 0–4), an improvement of one point which was found to be statistically significant on a Wilcoxon signed-rank test (p = 0.009). The median mJOA score improved by 4 points to 14 post-operatively (range 10–17) and this was also significant (p = 0.005). Spearman’s rho found no
EMG analysis of the timing of muscle activation showed a significant increase in duration of activation of TA from 37 % gait cycle duration pre-operatively to 41.7 % postoperatively (p = 0.02). Other muscles showed no
123
Changes in secondary outcome measures
Eur Spine J Table 2 Changes in kinematic key points following surgery Variable
Pre-op Mean
Post-op SD
Mean
Difference SD
Mean
95 % confidence intervals
p value
SD
Pelvic obliquity range
6.17
3.49
6.84
3.04
0.67
2.11
-0.60
1.95
0.27
Average pelvic tilt
8.76
5.43
8.10
6.48
-0.67
4.79
-3.56
2.23
0.63
Peak hip extension
-12.24
7.10
-13.10
7.34
-0.85
5.68
-4.29
2.58
0.6
Hip total sagittal plane excursion
42.46
7.62
42.52
8.03
0.07
5.48
-3.24
3.38
0.97
Hip abduction adduction range
12.00
3.47
11.07
4.07
-0.93
3.00
-2.74
0.88
0.29
Peak knee flexion in stance
14.82
5.83
16.20
6.27
1.39
1.28
-1.40
4.17
0.3
Peak knee flexion in swing Knee total sagittal plane excursion
47.80 49.59
1.73 8.67
50.07 50.37
2.21 8.94
2.26 0.79
1.40 3.49
-0.79 -1.32
5.32 2.89
0.13 0.43
Peak ankle dorsiflexion in stance
15.08
2.73
15.79
2.53
0.71
2.54
-16.73
13.42
0.33
Peak ankle dorsiflexion in swing
8.31
6.14
10.34
4.41
2.03
5.16
-1.08
5.15
0.18
-10.50
7.36
-9.86
5.25
0.64
4.37
-2.00
3.28
0.61
Peak ankle plantarflexion
Means, standard deviations (SD) and confidence intervals are reported in degrees Hip extension and ankle plantarflexion are negative in sign by convention
correlation between baseline mJOA and change in gait speed (q = 0.102, p = 0.74), or between change in mJOA and change in gait speed (q = 0.1105, p = 0.72).
Discussion The aim of surgical intervention in CSM is to halt the progression of neurological deterioration by decompressing the spinal cord and stabilising the affected spondylotic cervical levels [1]. It is generally accepted that CSM is typically progressive [17], although its precise natural history has not been established with any certainty. A recent 10-year prospective study comparing surgical and conservative management in 47 patients did not find inferior outcomes in the conservatively managed group, suggesting that deterioration is not inevitable [18]. Nonetheless, a lack of improvement in gait following surgery is not necessarily a negative outcome, but instead implies that the goal to stabilise the deficit and thereby prevent deterioration has been achieved. Our results found trends towards improvement in temporal–spatial and kinematic gait parameters, and significant improvements in important kinetics, as well as improved Nurick and mJOA scores. Peak ankle plantarflexor moment, power generation at the ankle in pre-swing and power absorption by the knee in terminal swing were all significantly higher post-operatively. Essentially, the locomotor system became more adept at absorbing and generating power at the knee and ankle at appropriate times during gait, in an overall pattern closer to that of agematched controls [4]. Stabilisation of the neurological deficit with surgery may have provided a window of opportunity to improve locomotion, unlike before surgery,
when the deficit was progressive and difficult to accommodate. That these kinetic improvements did not translate into faster gait speed is surprising, but may be explained by the fact that the baseline speed of the cohort was 1.05 m/s which, although slower than age-matched controls [4], is adequate for community ambulation. The quality rather than the speed of gait was the area of improvement. Other studies in CSM reported much lower baseline speeds, such as 0.56 m/s [19], and may have had greater scope for change. There is a possibility that the 0.03-m/s increase in speed could reflect the beginning of further increases after the twelfth post-operative month. Singh et al. [19] assessed performance on a 30-m timed walk test at 6, 12, 24 and 36 months post-surgery in CSM, and found that most improvement in gait speed occurred in the first 6 months. Similarly, a recent prospective study found the greatest increase in Berg Balance Scale scores in people with CSM in the first six post-operative months [20]. It would therefore seem that, while further improvement beyond the first post-operative year may be possible, the first 6–12 months are an important time to expect and evaluate changes in neurological function and mobility. An abnormal gait pattern in a person with neurological impairment reflects the direct consequences of the primary CNS lesion and the secondary compensatory processes that determine the optimal gait pattern for a given CNS lesion [21]. In the current study, it was possible that fear of falling or perceived lack of stability may have caused the CSM participants to retain a gait pattern characterised by slower gait speed with shorter stride lengths, less time in single support duration and smaller kinematic joint excursions, even though kinetic data showed improvements in the underlying strategies that produced this pattern. Similar
123
Eur Spine J Fig. 2 Kinetic data for hip, knee and ankle before and after surgery (significant differences are shown using arrows)
findings have been reported in people with hip osteoarthritis, who have reported relief of pain and improved quality of life following total hip arthroplasty, but without changes in gait kinematics and kinetics of either the affected or contralateral lower limb [22]. This lack of change in gait has been attributed to the preservation of the pain-avoidance strategies that characterised gait prior to joint replacement [22]. EMG data on the neuromuscular control of force production yielded somewhat conflicting results. The significant increase in the duration of activation of TA reflected a trend away from normal based on a previous study comparing EMG in gait in CSM and healthy controls [5]. However, as prolonged duration occurred during stance, it may have been a compensatory strategy to improve control around the ankle during single limb support, creating a more stable ankle joint. This could then facilitate a higher plantarflexor moment and power burst at pre-swing, leading to greater power absorption by the knee during swing, as shown by kinetics. Such increases in the duration of
123
muscle activity have been noted as compensatory strategies in the non-paretic lower limb after stroke [23]. This is the first study to provide quantitative data on the biomechanical and neuromuscular changes in gait following surgery. This new information may assist surgeons in effectively counselling people with CSM who seek greater clarity on the balance of risks and benefits in advance of a decision to undergo surgical intervention. The results of the current study hold important implications for the rehabilitation of gait in people with CSM following surgery. The improvements in kinetics and changes in EMG parameters suggested that some adaptation had taken place over the post-operative time period, although these did not translate into temporal–spatial or kinematic changes. Even if surgery does not directly alter the CNS lesion, the capacity for improvement in gait by neuroplasticity and compensation may be preserved [24]. Studies using functional MRI in CSM have provided evidence of reorganisation in the sensorimotor cortices following surgical decompression of cervical spine stenosis
Eur Spine J Table 3 Kinetic key points before and after surgery Variable
Pre-op Mean
Post-op SD
Mean
Difference SD
Mean
95 % confidence intervals
p value
*0.07
SD
Vertical GRF, peak 1
10.50
0.82
10.83
0.78
0.33
0.87
*
*
Antero-posterior GRF, acceleration
-1.48
0.62
-1.57
0.67
-0.09
0.28
-0.26
0.08
0.29
0.72
0.24
0.81
0.24
0.09
0.21
-0.04
0.21
0.15
Peak hip extensor moment Peak hip abductor moment
0.85
0.17
0.86
0.28
0.01
0.22
-0.12
0.15
0.82
Peak knee extensor moment
0.27
0.19
0.32
0.22
0.05
0.17
-0.06
0.16
0.32
Peak ankle plantarflexor moment
1.34
0.28
1.44
0.36
0.10
0.23
*
*
0.48 -0.56
0.26 0.32
0.42 -0.63
0.34 0.32
-0.06 -0.07
0.20 0.32
-0.18 -0.27
0.06 0.12
Hip power generation, loading response Hip power absorption, pre-swing Hip power generation, terminal swing Knee power absorption, loading response
*0.05 0.29 0.44
0.89
0.48
1.24
0.95
0.35
0.70
-0.08
0.77
0.099
-0.69
0.95
-0.93
1.11
-0.24
0.57
-0.59
0.10
0.15
Knee power absorption, pre-swing
-0.61
0.43
-0.91
0.77
-0.31
0.62
-0.68
0.07
0.098
Knee power absorption, terminal swing
-0.66
0.28
-1.08
0.55
-0.42
0.52
-0.73
-0.10
0.01
2.63
1.58
3.18
1.61
0.55
0.80
0.07
1.04
0.03
Ankle power generation, pre-swing
Ground reaction forces (GRF) are reported in newtons per kilogram, moments in newton metres per kilogram and powers in watts per kilogram Power absorption is negative in sign by convention The symbol * denotes a non-normally distributed variable with p value calculated using Wilcoxon signed-rank test
Table 4 Duration of muscle activation and co-activation as a percentage of gait cycle time before and after surgery Muscle
Pre-op
Post-op
Difference
95 % confidence interval
p value
Mean
SD
Mean
SD
Mean
SD
Rectus femoris
33.97
19.11
29.99
16.14
-3.97
18.63
-15.23
7.28
0.46
Biceps femoris
32.15
11.97
31.81
10.13
-0.34
13.77
-8.66
7.98
0.93
Tibialis anterior
36.99
11.16
41.67
10.83
4.68
6.46
0.78
8.58
0.02
Medial gastrocnemius
26.90
10.83
28.30
6.09
1.40
7.42
-3.08
5.89
0.51
RF/BF co-activation
13.64
8.69
12.71
10.90
-0.93
8.99
-6.36
4.50
TA/MG co-activation
3.97
3.79
5.83
5.65
1.85
5.00
*
*
0.72 *0.22
Duration of activation is expressed as a percentage of gait cycle duration RF rectus femoris, BF biceps femoris, TA tibialis anterior, MG medial gastrocnemius * Denotes variable tested with non-parametric Wilcoxon signed-rank test
[25]. The current study’s findings of improvement in aspects of gait are in keeping with these studies’ evidence of adaptation within the CNS following surgery. In a systematic review, Kokotilo [26] described the reorganisation of brain function in people with CNS damage as one of the fundamental mechanisms involved in recovery of sensorimotor function and commented that the brain networks involved in different aspects of motor control remain responsive, even in chronic paralysis. The findings of the current study add weight to the theoretical evidence that improvement in gait is possible following surgical decompression of CSM, and that post-operative rehabilitation should be pursued with this in mind. Our study has some limitations including the relatively small sample size, which, although statistically powered,
may have been too low to detect significance in smaller changes. People with CSM who were unable to ambulate 10 m prior to surgery had to be excluded as they could not complete a gait analysis protocol, and as such any changes in this more severely impaired population were not captured by this study. Finally, although a 12-month follow-up can be justified based on previous research, a longer-term post-operative analysis would evaluate changes beyond this time point.
Conclusion Three-dimensional gait analysis identified significant improvements in kinetic gait parameters 12 months
123
Eur Spine J Table 5 Amplitude of muscle activity bursts during the gait cycle before and after surgery Muscle burst
Pre-op Mean
Post-op SD
Difference
Mean
SD
Mean
Confidence interval
p value
SD
RF loading response
56.40
8.87
56.31
11.87
-0.09
7.67
-4.72
4.55
RF pre-swing
52.10
11.88
44.79
13.67
-7.31
14.99
-16.83
2.22
0.97 0.12
RF swing
17.46
12.59
21.31
9.06
3.85
11.60
-3.16
10.86
0.25
RF baseline
18.18
6.64
17.03
4.83
-1.15
6.01
-4.79
2.48
0.5
BF stance
43.76
9.22
45.21
8.26
1.46
10.77
-5.05
7.97
0.64
BF swing
46.13
10.12
44.39
16.87
-1.74
19.30
-13.40
9.92
0.75
BF baseline TA stance
16.81 55.46
3.95 8.89
15.17 55.01
3.93 9.33
-1.64 -0.45
4.82 4.30
-4.55 -3.05
1.28 2.15
0.24 0.71
TA swing
48.73
8.50
48.99
4.69
0.26
7.15
-4.07
4.58
0.9
TA baseline
15.42
4.28
16.84
3.57
1.41
3.69
-0.82
3.64
0.19
MG stance
59.19
4.39
57.27
4.90
-1.92
6.87
-6.07
2.23
0.33
MG baseline
12.40
3.77
11.93
3.71
-0.47
3.90
-2.83
1.88
0.67
Amplitude of each muscle is expressed as a percentage of its peak root-mean-square (RMS) amplitude during gait RF rectus femoris, BF biceps femoris, TA tibialis anterior, MG medial gastrocnemius, SD standard deviation
following surgical decompression in people with CSM. This was associated with electromyographic evidence of compensatory changes in stance and trends, albeit nonsignificant, towards improvement in gait speed and joint kinematics. These findings add to the growing body of evidence on surgical outcomes for CSM [2] and suggest the potential for further improvements through post-operative neuro-rehabilitation. Acknowledgments This study was funded by the Health Research Board of Ireland under Grant Number CTPF/2008/2. Conflict of interest
None.
References 1. Rao RD, Gourab K, David KS (2006) Operative treatment of cervical spondylotic myelopathy. J Bone Joint Surg Am 88(7):1619–1640 2. Fehlings MG, Wilson JR, Kopjar B, Yoon ST, Arnold PM, Massicotte EM, Vaccaro AR, Brodke DS, Shaffrey CI, Smith JS, Woodard EJ, Banco RJ, Chapman JR, Janssen ME, Bono CM, Sasso RC, Dekutoski MB, Gokaslan ZL (2013) Efficacy and safety of surgical decompression in patients with cervical spondylotic myelopathy: results of the AOSpine North America Prospective Multi-Center Study. J Bone Joint Surg Am 95(18):1651–1658 3. Rhee JM, Shamji MF, Erwin WM, Bransford RJ, Yoon ST, Smith JS, Kim HJ, Ely CG, Dettori JR, Patel AA, Kalsi-Ryan S (2013) Nonoperative management of cervical myelopathy: a systematic review. Spine 38(22S):S55–S67 4. Malone A, Meldrum D, Bolger C (2012) Gait impairment in cervical spondylotic myelopathy: comparison with age- and gender-matched healthy controls. Eur Spine J 21(12):2456–2466 5. Malone A, Meldrum D, Gleeson J, Bolger C (2013) Electromyographic characteristics of gait impairment in cervical spondylotic myelopathy. Eur Spine J 22(11):2538–2544
123
6. Kuhtz-Buschbeck JP, Johnk K, Muder S, Stolze H, Mehdorn M (1999) Analysis of gait in cervical myelopathy. Gait Posture 9(3):184–189 7. Maezawa Y, Uchida K, Baba H (2001) Gait analysis of spastic walking in patients with cervical compressive myelopathy. J Orthop Sci: Off J Japan Orthop Assoc 6(5):378–384 8. Davis R, Ounpuu S, Tyburski D, Gage J (1991) A gait analysis data collection and reduction technique. Hum Mov Sci 10:575–587 9. Hermens HJ, Freriks B (1997) The state of the art on sensors and sensor placement procedures for surface electromyography: a proposal for sensor placement procedures. Deliverable of the SENIAM project. Roessingh Research and Development, Enschede 10. Merletti R, Botter A, Troiano A, Merlo E, Minetto MA (2009) Technology and instrumentation for detection and conditioning of the surface electromyographic signal: state of the art. Clin Biomech (Bristol, Avon) 24(2):122–134 11. Malone A, Meldrum D, Gleeson J, Bolger C (2011) Reliability of surface electromyography timing parameters in gait in cervical spondylotic myelopathy. J Electromyogr Kinesiol 21(6):1004–1010 12. Mercer VS, Gross MT, Sharma S, Weeks E (2009) Comparison of gluteus medius muscle electromyographic activity during forward and lateral step-up exercises in older adults. Phys Ther 89(11):1205–1214 13. Lehman GJ, McGill SM (1999) The importance of normalization in the interpretation of surface electromyography: a proof of principle. J Manipulative Physiol Ther 22(7):444–446 14. Burden AM, Trew M, Baltzopoulos V (2003) Normalisation of gait EMGs: a re-examination. J Electromyogr Kinesiol 13(6):519–532 15. Judge JO, Schechtman K, Cress E (1996) The relationship between physical performance measures and independence in instrumental activities of daily living. J Am Geriatr Soc 44(11):1332–1341 16. McDermott A, Bolger C, Keating L, McEvoy L, Meldrum D (2010) Reliability of three-dimensional gait analysis in cervical spondylotic myelopathy. Gait Posture 32(4):552–558 17. Fehlings MG, Tetreault LA, Wilson JR, Skelly AC (2013) Cervical spondylotic myelopathy: current state of the art and future directions. Spine (Phila Pa 1976) 38(22 Suppl 1):S1–S8
Eur Spine J 18. Kadanka Z, Bednarik J, Novotny O, Urbanek I, Dusek L (2011) Cervical spondylotic myelopathy: conservative versus surgical treatment after 10 years. Eur Spine J 20(9):1533–1538 19. Singh A, Choi D, Crockard A (2009) Use of walking data in assessing operative results for cervical spondylotic myelopathy: long-term follow-up and comparison with controls. Spine (Phila Pa 1976) 34(12):1296–1300 20. Furlan JC, Kalsi-Ryan S, Kailaya-Vasan A, Massicotte EM, Fehlings MG (2011) Functional and clinical outcomes following surgical treatment in patients with cervical spondylotic myelopathy: a prospective study of 81 cases. J Neurosurg Spine 14(3):348–355 21. Dietz V (2002) Proprioception and locomotor disorders. Nat Rev Neurosci 3(10):781–790 22. Beaulieu ML, Lamontagne M, Beaule´ PE (2010) Lower limb biomechanics during gait do not return to normal following total hip arthroplasty. Gait Posture 32(2):269–273
23. Lamontagne A, Richards CL, Malouin F (2000) Coactivation during gait as an adaptive behavior after stroke. J Electromyogr Kinesiol 10(6):407–415 24. Dietz V, Harkema SJ (2004) Locomotor activity in spinal cordinjured persons. J Appl Physiol 96(5):1954–1960 25. Duggal N, Rabin D, Bartha R, Barry RL, Gati JS, Kowalczyk I, Fink M (2010) Brain reorganization in patients with spinal cord compression evaluated using fMRI. Neurology 74(13):1048–1054 26. Kokotilo KJ, Eng JJ, Curt A (2009) Reorganization and preservation of motor control of the brain in spinal cord injury: a systematic review. J Neurotrauma 26(11):2113–2126
123
ORIGINAL ARTICLE
Three-dimensional gait analysis outcomes at 1 year following decompressive surgery for cervical spondylotic myelopathy Ailish Malone • Dara Meldrum • Ciaran Bolger
Received: 6 December 2013 / Revised: 25 February 2014 / Accepted: 25 February 2014 Ó Springer-Verlag Berlin Heidelberg 2014
Abstract Purpose Gait impairment is an important feature of cervical sponydylotic myelopathy (CSM) as it can have a detrimental effect on function and quality of life. The aim of this study was to measure changes in gait in people with CSM following surgical decompression. Methods Thirteen participants with clinical and radiological evidence of CSM underwent three-dimensional gait analysis, using a full lower limb kinematic, kinetic and electromyography protocol, before and 12 months after decompressive surgery. Results No significant post-operative changes were detected in temporal–spatial or kinematic parameters. Kinetic data showed significant improvements in knee power absorption [mean improvement, 0.42 watts per kilogram (W/kg)], ankle plantarflexor moment (0.1 Nm/ kg) and ankle power generation (0.55 W/kg). Electromyography showed a 4.7 % increase in tibialis anterior activation time. Conclusions These findings indicate that improvement in locomotor function can be achieved after surgery. Future
A. Malone (&) Gait Analysis Laboratory, Central Remedial Clinic, Clontarf, Dublin 3, Ireland e-mail: [email protected] A. Malone D. Meldrum School of Physiotherapy, Royal College of Surgeons in Ireland, 123 St. Stephen’s Green, Dublin 2, Ireland e-mail: [email protected] C. Bolger Department of Neurosurgery, Beaumont Hospital, Dublin 9, Ireland e-mail: [email protected]
studies should explore the potential for further recovery of gait through targeted neuro-rehabilitation. Keywords Cervical myelopathy Gait analysis Electromyography Biomechanics Surgery
Introduction Surgery is widely used as a treatment for cervical spondylotic myelopathy (CSM), particularly in people whose symptoms show neurological progression over time. The aim of surgical intervention is to stabilise the disease by minimising spinal cord compression at the spondylotic levels, thereby preventing further neurological deterioration [1]. A recent prospective study found significant improvements in function, disability and quality of life 1 year after decompressive surgery for CSM [2]. However, surgery itself carries numerous risks including dysphagia, instrumentation failure or malposition, infection, and damage to neural structures, with an overall prevalence of adverse events of 18.7 % [2]. For many people with CSM, the decision to undergo surgery can be a difficult one and must be weighed against the possibility of further irreversible neurological deterioration if the cord is not decompressed [3]. Gait impairment is one of the cardinal symptoms of CSM. Our previous work determined the key kinematic, kinetic and electromyographic features that characterise gait in people with CSM compared to age- and gendermatched healthy controls [4, 5]. There is some evidence from follow-up studies that gait speed can improve after surgery by up to 0.15 m/s [6, 7]. A more detailed examination of gait before and after surgical intervention is required to assess the true potential for recovery, as well as
123
Eur Spine J
Methods
to-noise ratio of at least 50 dB, amplified with a gain range of 2,000–13,200 and sampled into a PC at 1,000 Hz using a 32-channel DI-720 analogue to digital convertor with 12-bit resolution (DATAQ Instruments Inc., Akron, Ohio, USA). The assessment consisted of a static trial for calibration and a warm-up trial, followed by a number of barefoot walking trials at self-selected speed along a 12-m overground walkway. The assessment continued until ten trials with good-quality data, including five left and right force plate strikes, had been achieved. Rest breaks were provided as required to avoid fatigue.
Recruitment
Data processing
People with CSM were recruited from a neurosurgical spinal assessment clinic at a national neurosurgical hospital over a 2-year period from December 2008 to December 2010. The following inclusion criteria were applied: (1) aged 18 years or over; (2) able to give informed consent; (3) able to mobilise to at least 10 m without assistance of another person (but using aids if necessary); (4) clinical and radiological evidence of CSM. Patients were excluded if they were affected by any of the following: (1) severe respiratory or cardiac disease hindering safe mobilisation; (2) history of neurological disorders with persistent deficit; (3) symptomatic musculoskeletal problems affecting gait; (4) tandem lumbar spine stenosis; (5) previous surgical decompression for CSM. The severity of myelopathy was assessed using the Nurick scale and modified Japanese Orthopaedic Association (mJOA) score.
Gait data were processed using VICON’s Plug-in GaitÒ software. The average of ten trials was used to represent the gait pattern of each participant pre- and post-surgery. Kinematic data were derived for pelvis, hip, knee and ankle in three planes. Kinetic data for joint moments and powers were derived for the hip in the sagittal and coronal planes, and for the knee and ankle in the sagittal plane. EMG signals from each gait trial were imported into MATLAB (The Mathworks Inc., Natick, MA, USA) and filtered with a fourth-order Butterworth low-pass filter of 400 Hz and a second-order Butterworth high-pass filter of 25 Hz, applied in forward and reverse directions to remove motion artefact [10]. Muscle activity was detected using an algorithm, previously validated in this population, which identifies changes in amplitude and frequency from its baseline non-active state to determine contraction [11]. Duration of muscle activity and co-activation between agonist and antagonist muscles was reported as a percentage of gait cycle duration. The EMG amplitude of each burst of muscle activity was extracted by calculating the root-mean square (RMS) of the signal over a 30-ms window with 20-ms overlap [12]. Signals were normalised by expressing the RMS amplitude at each time point as a percentage of the maximum amplitude measured for that muscle during gait [13, 14].
the clinical significance of that recovery. This could provide further information for surgeons and patients on the expected outcomes of surgery and contribute to the expanding base of knowledge on the potential for changes in gait following spinal cord injury. The aim of this study was to determine changes in temporal–spatial, kinematic, kinetic and electromyography (EMG) variables in gait in people with CSM 1 year following surgery.
Gait analysis protocol Three-dimensional gait analysis (3DGA) was conducted prior to and 12 months after surgery using a VICONÒ250 five-camera Motion Analysis System (VICON, Oxford, UK) and an integrated Kistler multi-component force plate (Kistler, Winterthur, Switzerland). Kinematic and kinetic data were collected according to a standard protocol involving the placement of 15 reflective markers to anatomical landmarks on the lower limbs, tracked at 50 Hertz (Hz) by the VICON system [8]. Surface EMG signals were recorded during gait from rectus femoris (RF), biceps femoris (BF), tibialis anterior (TA) and medial head of gastrocnemius (MG), using double-differentiated pre-amplified stainless steel electrodes with a common mode rejection ratio of 100 decibels (dB) at 65 Hz. The skin underlying each electrode was shaved and cleaned to reduce impedance. Electrode placement followed the guidelines of SENIAM [9]. A reference electrode was placed over C7. Signals were collected across a bandwidth of 20–500 Hz with a signal-
123
Statistical analysis Reports have considered the minimal clinically important difference in gait speed to be 0.1 m/s as this shows a good correlation with the ability to carry out functional activities of daily living [15]. A power calculation based on a paired t test found that a sample size of 13 participants would have 90 % power at the 5 % significant level to detect a change in gait speed of 0.1 m/s from pre- to post-operative assessment, with a standard deviation of 0.11 m/s in line with previously published data [6]. Differences in gait parameters from pre- to post-surgery were tested using
Eur Spine J Table 1 Changes in temporal–spatial parameters following surgery Variable
Pre-op Mean
Gait speed (m/s) Cadence (steps/min) Stride length (m)
Post-op SD
Mean
Difference SD
Mean
95 % confidence intervals
p value
SD
1.05
0.32
1.08
0.37
0.03
0.20
-0.09
0.15
0.62
110.30
12.44
111.37
14.87
1.06
17.96
-9.79
11.9
0.83
1.12
0.26
1.14
0.31
0.02
0.11
-0.05
0.09
0.57
Double support duration (% GC)
27.99
6.90
28.72
8.93
0.72
4.88
*
*
Single support duration (% GC)
35.89
2.82
35.68
4.46
-0.22
3.27
-2.19
1.76
0.82
Opposite foot off (% GC)
13.38
2.78
13.79
3.88
0.41
2.83
-1.30
2.12
0.61
Opposite foot contact (% GC) Foot off (% GC)
49.27 63.87
1.31 4.57
49.48 64.39
1.38 4.66
0.21 0.52
1.25 2.20
-0.55 -0.81
0.96 1.85
0.56 0.41
0.16
0.04
0.17
0.05
0.01
0.02
-0.01
0.02
0.26
Step width (m)
*0.75
SD standard deviation, m metres, s seconds, min minute, GC gait cycle * Denotes non-normally distributed variable tested with Wilcoxon signed-rank test
paired t tests where data satisfied the assumption of normal distribution, and using Wilcoxon signed-rank tests for nonnormally distributed data. Analysis was conducted in Stata IC 11 (StataCorp, Texas, USA).
Results Participants Forty-four people met the inclusion criteria over the 24-month recruitment period. Sixteen were excluded due to immobility (four), previous surgery for CSM (four) and pre-existing neurological or orthopaedic conditions (eight). A further ten declined to participate and five did not proceed to surgery. The remaining 13 people with CSM underwent decompressive surgery and gave informed consent to participate. Ten participants were followed up at 12 months as per the study protocol and three participants attended a 6-month follow-up due to time limitations. The mean time from pre-operative assessment to surgery was 2.2 months (range 0–4.9 months). The mean age of the cohort was 56.6 years (range 34–77 years). Myelopathic symptoms were reported for a mean of 59 months prior to surgery (range 9–420 months). The median pre-operative Nurick score was 3 (range 1–4) and median mJOA score was 10 (range 8–13). The most commonly involved cervical levels were C3/ 4, C4/5 and C5/6. Seven participants had an anterior surgical approach, involving an anterior cervical discectomy and fusion (ACDF) in six and an anterior cervical corpectomy and fusion (ACCF) in one case. The remaining six participants had a posterior approach, three of which involved laminectomy and lateral mass plating (LLMP), one, a decompression and lateral mass screws, and two, an occipital–cervical fusion.
Two participants (15 %) suffered post-operative complications. One had a wound infection at the surgical site and subsequent sepsis, and the other experienced instrumentation failure. In both cases, the instrumentation was removed and the cervical spine stabilised in a halo ring and brace. The second participant then underwent a repeat fusion following the removal of the halo brace. These complication rates were in line with a recent large case series [2]. Changes in temporal–spatial parameters Gait speed increased from a mean of 1.05 m/s pre-operatively to 1.08 m/s at the 12-month post-operative assessment. This was not statistically significant (p = 0.62). Similarly, none of the measured temporal–spatial parameters changed significantly (Table 1). Changes in kinematics Kinematic data are shown in Fig. 1. Graphs indicated a post-operative decrease in anterior pelvic tilt, an increase in peak hip extension and an increase in peak knee flexion in swing. These changes were not statistically significant (Table 2). Changes in kinetics Joint moments and powers pre- and post-surgery are shown in Fig. 2. Peak ankle plantarflexor moment increased significantly from 1.34 Nm/kg pre-surgery to 1.44 Nm/kg post-surgery (p = 0.05). There was a significant increase in ankle power at pre-swing from 2.63 to 3.18 W/kg (p = 0.03) and knee power absorption at terminal swing, from 0.66 to 1.08 W/kg (p = 0.01). Knee power in initial swing showed a non-significant increase
123
Eur Spine J Fig. 1 Kinematic data for pelvis, hip, knee and ankle before and after surgery. HC healthy control data
from 0.69 to 0.93 W/kg (p = 0.099). Similarly, hip power at pre-swing increased from 0.89 to 1.24 W/kg (p = 0.098) (Table 3). All changes exceeded previously calculated standard errors of measurement for this population [16].
significant change in timing (Table 4). Similarly, there were no changes in the amplitude of muscle activity (Table 5).
Changes in electromyography
The median post-operative Nurick score was 2 (range 0–4), an improvement of one point which was found to be statistically significant on a Wilcoxon signed-rank test (p = 0.009). The median mJOA score improved by 4 points to 14 post-operatively (range 10–17) and this was also significant (p = 0.005). Spearman’s rho found no
EMG analysis of the timing of muscle activation showed a significant increase in duration of activation of TA from 37 % gait cycle duration pre-operatively to 41.7 % postoperatively (p = 0.02). Other muscles showed no
123
Changes in secondary outcome measures
Eur Spine J Table 2 Changes in kinematic key points following surgery Variable
Pre-op Mean
Post-op SD
Mean
Difference SD
Mean
95 % confidence intervals
p value
SD
Pelvic obliquity range
6.17
3.49
6.84
3.04
0.67
2.11
-0.60
1.95
0.27
Average pelvic tilt
8.76
5.43
8.10
6.48
-0.67
4.79
-3.56
2.23
0.63
Peak hip extension
-12.24
7.10
-13.10
7.34
-0.85
5.68
-4.29
2.58
0.6
Hip total sagittal plane excursion
42.46
7.62
42.52
8.03
0.07
5.48
-3.24
3.38
0.97
Hip abduction adduction range
12.00
3.47
11.07
4.07
-0.93
3.00
-2.74
0.88
0.29
Peak knee flexion in stance
14.82
5.83
16.20
6.27
1.39
1.28
-1.40
4.17
0.3
Peak knee flexion in swing Knee total sagittal plane excursion
47.80 49.59
1.73 8.67
50.07 50.37
2.21 8.94
2.26 0.79
1.40 3.49
-0.79 -1.32
5.32 2.89
0.13 0.43
Peak ankle dorsiflexion in stance
15.08
2.73
15.79
2.53
0.71
2.54
-16.73
13.42
0.33
Peak ankle dorsiflexion in swing
8.31
6.14
10.34
4.41
2.03
5.16
-1.08
5.15
0.18
-10.50
7.36
-9.86
5.25
0.64
4.37
-2.00
3.28
0.61
Peak ankle plantarflexion
Means, standard deviations (SD) and confidence intervals are reported in degrees Hip extension and ankle plantarflexion are negative in sign by convention
correlation between baseline mJOA and change in gait speed (q = 0.102, p = 0.74), or between change in mJOA and change in gait speed (q = 0.1105, p = 0.72).
Discussion The aim of surgical intervention in CSM is to halt the progression of neurological deterioration by decompressing the spinal cord and stabilising the affected spondylotic cervical levels [1]. It is generally accepted that CSM is typically progressive [17], although its precise natural history has not been established with any certainty. A recent 10-year prospective study comparing surgical and conservative management in 47 patients did not find inferior outcomes in the conservatively managed group, suggesting that deterioration is not inevitable [18]. Nonetheless, a lack of improvement in gait following surgery is not necessarily a negative outcome, but instead implies that the goal to stabilise the deficit and thereby prevent deterioration has been achieved. Our results found trends towards improvement in temporal–spatial and kinematic gait parameters, and significant improvements in important kinetics, as well as improved Nurick and mJOA scores. Peak ankle plantarflexor moment, power generation at the ankle in pre-swing and power absorption by the knee in terminal swing were all significantly higher post-operatively. Essentially, the locomotor system became more adept at absorbing and generating power at the knee and ankle at appropriate times during gait, in an overall pattern closer to that of agematched controls [4]. Stabilisation of the neurological deficit with surgery may have provided a window of opportunity to improve locomotion, unlike before surgery,
when the deficit was progressive and difficult to accommodate. That these kinetic improvements did not translate into faster gait speed is surprising, but may be explained by the fact that the baseline speed of the cohort was 1.05 m/s which, although slower than age-matched controls [4], is adequate for community ambulation. The quality rather than the speed of gait was the area of improvement. Other studies in CSM reported much lower baseline speeds, such as 0.56 m/s [19], and may have had greater scope for change. There is a possibility that the 0.03-m/s increase in speed could reflect the beginning of further increases after the twelfth post-operative month. Singh et al. [19] assessed performance on a 30-m timed walk test at 6, 12, 24 and 36 months post-surgery in CSM, and found that most improvement in gait speed occurred in the first 6 months. Similarly, a recent prospective study found the greatest increase in Berg Balance Scale scores in people with CSM in the first six post-operative months [20]. It would therefore seem that, while further improvement beyond the first post-operative year may be possible, the first 6–12 months are an important time to expect and evaluate changes in neurological function and mobility. An abnormal gait pattern in a person with neurological impairment reflects the direct consequences of the primary CNS lesion and the secondary compensatory processes that determine the optimal gait pattern for a given CNS lesion [21]. In the current study, it was possible that fear of falling or perceived lack of stability may have caused the CSM participants to retain a gait pattern characterised by slower gait speed with shorter stride lengths, less time in single support duration and smaller kinematic joint excursions, even though kinetic data showed improvements in the underlying strategies that produced this pattern. Similar
123
Eur Spine J Fig. 2 Kinetic data for hip, knee and ankle before and after surgery (significant differences are shown using arrows)
findings have been reported in people with hip osteoarthritis, who have reported relief of pain and improved quality of life following total hip arthroplasty, but without changes in gait kinematics and kinetics of either the affected or contralateral lower limb [22]. This lack of change in gait has been attributed to the preservation of the pain-avoidance strategies that characterised gait prior to joint replacement [22]. EMG data on the neuromuscular control of force production yielded somewhat conflicting results. The significant increase in the duration of activation of TA reflected a trend away from normal based on a previous study comparing EMG in gait in CSM and healthy controls [5]. However, as prolonged duration occurred during stance, it may have been a compensatory strategy to improve control around the ankle during single limb support, creating a more stable ankle joint. This could then facilitate a higher plantarflexor moment and power burst at pre-swing, leading to greater power absorption by the knee during swing, as shown by kinetics. Such increases in the duration of
123
muscle activity have been noted as compensatory strategies in the non-paretic lower limb after stroke [23]. This is the first study to provide quantitative data on the biomechanical and neuromuscular changes in gait following surgery. This new information may assist surgeons in effectively counselling people with CSM who seek greater clarity on the balance of risks and benefits in advance of a decision to undergo surgical intervention. The results of the current study hold important implications for the rehabilitation of gait in people with CSM following surgery. The improvements in kinetics and changes in EMG parameters suggested that some adaptation had taken place over the post-operative time period, although these did not translate into temporal–spatial or kinematic changes. Even if surgery does not directly alter the CNS lesion, the capacity for improvement in gait by neuroplasticity and compensation may be preserved [24]. Studies using functional MRI in CSM have provided evidence of reorganisation in the sensorimotor cortices following surgical decompression of cervical spine stenosis
Eur Spine J Table 3 Kinetic key points before and after surgery Variable
Pre-op Mean
Post-op SD
Mean
Difference SD
Mean
95 % confidence intervals
p value
*0.07
SD
Vertical GRF, peak 1
10.50
0.82
10.83
0.78
0.33
0.87
*
*
Antero-posterior GRF, acceleration
-1.48
0.62
-1.57
0.67
-0.09
0.28
-0.26
0.08
0.29
0.72
0.24
0.81
0.24
0.09
0.21
-0.04
0.21
0.15
Peak hip extensor moment Peak hip abductor moment
0.85
0.17
0.86
0.28
0.01
0.22
-0.12
0.15
0.82
Peak knee extensor moment
0.27
0.19
0.32
0.22
0.05
0.17
-0.06
0.16
0.32
Peak ankle plantarflexor moment
1.34
0.28
1.44
0.36
0.10
0.23
*
*
0.48 -0.56
0.26 0.32
0.42 -0.63
0.34 0.32
-0.06 -0.07
0.20 0.32
-0.18 -0.27
0.06 0.12
Hip power generation, loading response Hip power absorption, pre-swing Hip power generation, terminal swing Knee power absorption, loading response
*0.05 0.29 0.44
0.89
0.48
1.24
0.95
0.35
0.70
-0.08
0.77
0.099
-0.69
0.95
-0.93
1.11
-0.24
0.57
-0.59
0.10
0.15
Knee power absorption, pre-swing
-0.61
0.43
-0.91
0.77
-0.31
0.62
-0.68
0.07
0.098
Knee power absorption, terminal swing
-0.66
0.28
-1.08
0.55
-0.42
0.52
-0.73
-0.10
0.01
2.63
1.58
3.18
1.61
0.55
0.80
0.07
1.04
0.03
Ankle power generation, pre-swing
Ground reaction forces (GRF) are reported in newtons per kilogram, moments in newton metres per kilogram and powers in watts per kilogram Power absorption is negative in sign by convention The symbol * denotes a non-normally distributed variable with p value calculated using Wilcoxon signed-rank test
Table 4 Duration of muscle activation and co-activation as a percentage of gait cycle time before and after surgery Muscle
Pre-op
Post-op
Difference
95 % confidence interval
p value
Mean
SD
Mean
SD
Mean
SD
Rectus femoris
33.97
19.11
29.99
16.14
-3.97
18.63
-15.23
7.28
0.46
Biceps femoris
32.15
11.97
31.81
10.13
-0.34
13.77
-8.66
7.98
0.93
Tibialis anterior
36.99
11.16
41.67
10.83
4.68
6.46
0.78
8.58
0.02
Medial gastrocnemius
26.90
10.83
28.30
6.09
1.40
7.42
-3.08
5.89
0.51
RF/BF co-activation
13.64
8.69
12.71
10.90
-0.93
8.99
-6.36
4.50
TA/MG co-activation
3.97
3.79
5.83
5.65
1.85
5.00
*
*
0.72 *0.22
Duration of activation is expressed as a percentage of gait cycle duration RF rectus femoris, BF biceps femoris, TA tibialis anterior, MG medial gastrocnemius * Denotes variable tested with non-parametric Wilcoxon signed-rank test
[25]. The current study’s findings of improvement in aspects of gait are in keeping with these studies’ evidence of adaptation within the CNS following surgery. In a systematic review, Kokotilo [26] described the reorganisation of brain function in people with CNS damage as one of the fundamental mechanisms involved in recovery of sensorimotor function and commented that the brain networks involved in different aspects of motor control remain responsive, even in chronic paralysis. The findings of the current study add weight to the theoretical evidence that improvement in gait is possible following surgical decompression of CSM, and that post-operative rehabilitation should be pursued with this in mind. Our study has some limitations including the relatively small sample size, which, although statistically powered,
may have been too low to detect significance in smaller changes. People with CSM who were unable to ambulate 10 m prior to surgery had to be excluded as they could not complete a gait analysis protocol, and as such any changes in this more severely impaired population were not captured by this study. Finally, although a 12-month follow-up can be justified based on previous research, a longer-term post-operative analysis would evaluate changes beyond this time point.
Conclusion Three-dimensional gait analysis identified significant improvements in kinetic gait parameters 12 months
123
Eur Spine J Table 5 Amplitude of muscle activity bursts during the gait cycle before and after surgery Muscle burst
Pre-op Mean
Post-op SD
Difference
Mean
SD
Mean
Confidence interval
p value
SD
RF loading response
56.40
8.87
56.31
11.87
-0.09
7.67
-4.72
4.55
RF pre-swing
52.10
11.88
44.79
13.67
-7.31
14.99
-16.83
2.22
0.97 0.12
RF swing
17.46
12.59
21.31
9.06
3.85
11.60
-3.16
10.86
0.25
RF baseline
18.18
6.64
17.03
4.83
-1.15
6.01
-4.79
2.48
0.5
BF stance
43.76
9.22
45.21
8.26
1.46
10.77
-5.05
7.97
0.64
BF swing
46.13
10.12
44.39
16.87
-1.74
19.30
-13.40
9.92
0.75
BF baseline TA stance
16.81 55.46
3.95 8.89
15.17 55.01
3.93 9.33
-1.64 -0.45
4.82 4.30
-4.55 -3.05
1.28 2.15
0.24 0.71
TA swing
48.73
8.50
48.99
4.69
0.26
7.15
-4.07
4.58
0.9
TA baseline
15.42
4.28
16.84
3.57
1.41
3.69
-0.82
3.64
0.19
MG stance
59.19
4.39
57.27
4.90
-1.92
6.87
-6.07
2.23
0.33
MG baseline
12.40
3.77
11.93
3.71
-0.47
3.90
-2.83
1.88
0.67
Amplitude of each muscle is expressed as a percentage of its peak root-mean-square (RMS) amplitude during gait RF rectus femoris, BF biceps femoris, TA tibialis anterior, MG medial gastrocnemius, SD standard deviation
following surgical decompression in people with CSM. This was associated with electromyographic evidence of compensatory changes in stance and trends, albeit nonsignificant, towards improvement in gait speed and joint kinematics. These findings add to the growing body of evidence on surgical outcomes for CSM [2] and suggest the potential for further improvements through post-operative neuro-rehabilitation. Acknowledgments This study was funded by the Health Research Board of Ireland under Grant Number CTPF/2008/2. Conflict of interest
None.
References 1. Rao RD, Gourab K, David KS (2006) Operative treatment of cervical spondylotic myelopathy. J Bone Joint Surg Am 88(7):1619–1640 2. Fehlings MG, Wilson JR, Kopjar B, Yoon ST, Arnold PM, Massicotte EM, Vaccaro AR, Brodke DS, Shaffrey CI, Smith JS, Woodard EJ, Banco RJ, Chapman JR, Janssen ME, Bono CM, Sasso RC, Dekutoski MB, Gokaslan ZL (2013) Efficacy and safety of surgical decompression in patients with cervical spondylotic myelopathy: results of the AOSpine North America Prospective Multi-Center Study. J Bone Joint Surg Am 95(18):1651–1658 3. Rhee JM, Shamji MF, Erwin WM, Bransford RJ, Yoon ST, Smith JS, Kim HJ, Ely CG, Dettori JR, Patel AA, Kalsi-Ryan S (2013) Nonoperative management of cervical myelopathy: a systematic review. Spine 38(22S):S55–S67 4. Malone A, Meldrum D, Bolger C (2012) Gait impairment in cervical spondylotic myelopathy: comparison with age- and gender-matched healthy controls. Eur Spine J 21(12):2456–2466 5. Malone A, Meldrum D, Gleeson J, Bolger C (2013) Electromyographic characteristics of gait impairment in cervical spondylotic myelopathy. Eur Spine J 22(11):2538–2544
123
6. Kuhtz-Buschbeck JP, Johnk K, Muder S, Stolze H, Mehdorn M (1999) Analysis of gait in cervical myelopathy. Gait Posture 9(3):184–189 7. Maezawa Y, Uchida K, Baba H (2001) Gait analysis of spastic walking in patients with cervical compressive myelopathy. J Orthop Sci: Off J Japan Orthop Assoc 6(5):378–384 8. Davis R, Ounpuu S, Tyburski D, Gage J (1991) A gait analysis data collection and reduction technique. Hum Mov Sci 10:575–587 9. Hermens HJ, Freriks B (1997) The state of the art on sensors and sensor placement procedures for surface electromyography: a proposal for sensor placement procedures. Deliverable of the SENIAM project. Roessingh Research and Development, Enschede 10. Merletti R, Botter A, Troiano A, Merlo E, Minetto MA (2009) Technology and instrumentation for detection and conditioning of the surface electromyographic signal: state of the art. Clin Biomech (Bristol, Avon) 24(2):122–134 11. Malone A, Meldrum D, Gleeson J, Bolger C (2011) Reliability of surface electromyography timing parameters in gait in cervical spondylotic myelopathy. J Electromyogr Kinesiol 21(6):1004–1010 12. Mercer VS, Gross MT, Sharma S, Weeks E (2009) Comparison of gluteus medius muscle electromyographic activity during forward and lateral step-up exercises in older adults. Phys Ther 89(11):1205–1214 13. Lehman GJ, McGill SM (1999) The importance of normalization in the interpretation of surface electromyography: a proof of principle. J Manipulative Physiol Ther 22(7):444–446 14. Burden AM, Trew M, Baltzopoulos V (2003) Normalisation of gait EMGs: a re-examination. J Electromyogr Kinesiol 13(6):519–532 15. Judge JO, Schechtman K, Cress E (1996) The relationship between physical performance measures and independence in instrumental activities of daily living. J Am Geriatr Soc 44(11):1332–1341 16. McDermott A, Bolger C, Keating L, McEvoy L, Meldrum D (2010) Reliability of three-dimensional gait analysis in cervical spondylotic myelopathy. Gait Posture 32(4):552–558 17. Fehlings MG, Tetreault LA, Wilson JR, Skelly AC (2013) Cervical spondylotic myelopathy: current state of the art and future directions. Spine (Phila Pa 1976) 38(22 Suppl 1):S1–S8
Eur Spine J 18. Kadanka Z, Bednarik J, Novotny O, Urbanek I, Dusek L (2011) Cervical spondylotic myelopathy: conservative versus surgical treatment after 10 years. Eur Spine J 20(9):1533–1538 19. Singh A, Choi D, Crockard A (2009) Use of walking data in assessing operative results for cervical spondylotic myelopathy: long-term follow-up and comparison with controls. Spine (Phila Pa 1976) 34(12):1296–1300 20. Furlan JC, Kalsi-Ryan S, Kailaya-Vasan A, Massicotte EM, Fehlings MG (2011) Functional and clinical outcomes following surgical treatment in patients with cervical spondylotic myelopathy: a prospective study of 81 cases. J Neurosurg Spine 14(3):348–355 21. Dietz V (2002) Proprioception and locomotor disorders. Nat Rev Neurosci 3(10):781–790 22. Beaulieu ML, Lamontagne M, Beaule´ PE (2010) Lower limb biomechanics during gait do not return to normal following total hip arthroplasty. Gait Posture 32(2):269–273
23. Lamontagne A, Richards CL, Malouin F (2000) Coactivation during gait as an adaptive behavior after stroke. J Electromyogr Kinesiol 10(6):407–415 24. Dietz V, Harkema SJ (2004) Locomotor activity in spinal cordinjured persons. J Appl Physiol 96(5):1954–1960 25. Duggal N, Rabin D, Bartha R, Barry RL, Gati JS, Kowalczyk I, Fink M (2010) Brain reorganization in patients with spinal cord compression evaluated using fMRI. Neurology 74(13):1048–1054 26. Kokotilo KJ, Eng JJ, Curt A (2009) Reorganization and preservation of motor control of the brain in spinal cord injury: a systematic review. J Neurotrauma 26(11):2113–2126
123
