Keywords
C5 palsy - cervical decompression - myelopathy
Introduction
Cervical spondylotic myelopathy affects the middle to elderly population, causing
significant morbidity. Keegan[1] first described the C5 nerve palsy that developed following posterior decompression
surgeries as a "dissociated motor loss." C5 palsy causes significant morbidity to
patients, making them unable to perform their daily activities.[2]
[3] Both anterior and posterior decompression surgeries can cause this complication.[4]
[5] Controversy still exists regarding the incidence, cause, risk factors, and measures
to prevent C5 palsy. Incidence varied between 0 to 26.4% in anterior surgeries and
0 to 50% in posterior surgeries.[6] Here, we tried to study the incidence in both anterior and posterior surgeries as
well as identify various risk factors that lead to development of C5 nerve palsy.
Patients and Methods
Retrospective analysis of all patient records who were admitted between the year 2006
and 2015 with a diagnosis of cervical compressive myelopathy was performed. All cases
included in this study had surgery involving C4–5 segment. Patients who underwent
either anterior or posterior surgery were included. Patients who underwent both anterior
and posterior surgeries, previous cervical surgeries, those with trauma, infection,
or inflammatory disorders were excluded.
Demographic and clinical features of each patient were recorded. These included age,
gender, diagnosis, preoperative Japanese Orthopedic Association (JOA) score and Nurick
grade, surgery performed, and final JOA score and Nurick grade. All patients who developed
weakness in power of deltoid and/or biceps by ≥ 1manual muscle testing (MMT) grade,
without involvement of other muscles or worsening of myelopathy were included in the
palsy group, rest all patients were included in the non-palsy group.
Preoperative and postoperative cervical Cobb angle (C2–C7) and C4–5 intervertebral
angle is obtained from lateral radiographs ([Figs. 1] and [2]). From axial computed tomography (CT) scan, C4–5 foramen width was measured at its
narrowest point.
Fig. 1 (A) Measurement of cervical lordosis and (B) C45 intervertebral angle.
Fig. 2 Width of C5 foramen, measured at narrowest point on axial computed tomography scan.
Two senior spine surgeons with a minimum experience of 15 years performed all the
surgeries and clinical assessments. Two spine residents, who were blinded to the C5
palsy status, performed all the measurements. No neuromonitoring was used and steel
burr with saline at room temperature was used in all cases.
In patients with C5 palsy, onset, duration, treatment given, and level of recovery
at final follow-up was noted.
Comparison was made between palsy and non-palsy groups to identify the factors that
would predict C5 palsy following decompression surgeries. Change in overall lordosis
(Cobb angle) and intervertebral angle was correlated for incidence of C5 palsy. C45
foramen width on the palsy side was compared with contralateral side, and the mean
width was compared between palsy and non-palsy groups.
Comparisons were also made between patients with ossification of the posterior longitudinal
ligament (OPLL) and all other patients as well as between anterior and posterior surgeries.
Results
From 2006 to 2015, we identified 421 patients who underwent surgery for cervical compressive
myelopathy, out of which 31 patients were excluded (214 were males and 176 were females).
A total of 72 patients had OPLL and rest 318 patients had cervical spondylotic myelopathy
(CSM).
In the CSM group, 205 patients underwent anterior surgery, while 113 patients underwent
posterior surgery and in the OPLL group, 27 patients underwent anterior surgery while
45 underwent posterior surgery. A total of 10 patients developed C5 palsy out of which
6 were males and 4 were females. Mean age of patients in palsy group was 55.9 years.
Out of these 10 patients, 6 had OPLL, while 4 had CSM. Characteristics of patients
in both palsy and non-palsy group are shown in [Table 1]. The details of patients with C5 palsy are shown in [Table 2].
Table 1
Characteristics of patients in palsy and non-palsy group
|
Palsy group
|
Non-palsy group
|
|
Age (mean)
|
55.9 y
|
62.3 y
|
|
Gender ratio (M:F)
|
1:1
|
1.2:1
|
|
Total no. of patients
|
10
|
380
|
|
Disease
|
|
OPLL
|
6
|
66
|
|
CSM
|
4
|
314
|
|
JOA (mean)
|
|
Pre-op
|
11.4
|
10.4
|
|
At final follow-up
|
13.3
|
12.3
|
|
Nurick grade (mean)
|
|
Pre-op
|
2.5
|
3.0
|
|
At final follow-up
|
1.7
|
2.2
|
|
Surgeries
|
|
ACDF
|
0
|
128
|
|
ACCF
|
0
|
104
|
|
Laminoplasty
|
2
|
42
|
|
Laminectomy
|
6
|
79
|
|
Laminectomy with fixation
|
2
|
27
|
Abbreviations: ACCF, anterior cervical corpectomy and fusion; ACDF, anterior cervical
discectomy and fusion; CSM, cervical spondylotic myelopathy; F, female; JOA, Japanese
Orthopedic Association; M, male; OPLL, ossification of the posterior longitudinal
ligament.
Table 2
Details of patients with postoperative C5 palsy
|
Age
|
Sex
|
Diagnosis
|
Treatment given
|
Nurick grade (pre)
|
Nurick grade (final)
|
Onset of palsy (days postop)
|
Grade of palsy at onset (MMT)
|
Grade of palsy at final follow-up
|
Time for recovery (mo)
|
Follow-up duration (mo)
|
|
87
|
Male
|
CSM
|
Laminoplasty
|
2
|
2
|
2
|
3
|
4
|
4
|
22
|
|
33
|
Male
|
OPLL
|
Laminectomy + fixation
|
5
|
3
|
3
|
4
|
5
|
3
|
12
|
|
60
|
Male
|
OPLL
|
Laminectomy
|
2
|
1
|
7
|
2
|
4
|
10
|
18
|
|
66
|
Male
|
CSM
|
Laminectomy
|
1
|
1
|
2
|
3
|
4
|
6
|
36
|
|
63
|
Female
|
OPLL
|
Laminectomy
|
4
|
2
|
1
|
4
|
5
|
2
|
24
|
|
52
|
Female
|
OPLL
|
Laminoplasty
|
2
|
1
|
1
|
1
|
2
|
|
48
|
|
39
|
Female
|
CSM
|
Laminectomy
|
3
|
2
|
3
|
3
|
4
|
8
|
26
|
|
61
|
Male
|
OPLL
|
Laminectomy
|
2
|
1
|
4
|
4
|
5
|
4
|
18
|
|
28
|
Female
|
CSM
|
Laminectomy
|
1
|
1
|
2
|
3
|
4
|
8
|
24
|
|
70
|
Female
|
OPLL
|
Laminectomy + fixation
|
3
|
3
|
3
|
2
|
4
|
12
|
16
|
Abbreviations: CSM, cervical spondylotic myelopathy; MMT, manual muscle testing, OPLL,
ossification of the posterior longitudinal ligament.
Analysis
Surgical procedure: A total of 232 patients underwent anterior surgery, out of which
128 had anterior cervical discectomy and fusion (ACDF) (53 were single level, 75 were
double level) and 104 patients had single-level anterior cervical corpectomy and fusion
(ACCF). Out of 232 patients who underwent anterior surgeries, 27 patients had OPLL
while rest had CSM. None of these patients developed palsy of C5 nerve.
Out of 158 patients who underwent posterior surgery, laminectomy was performed in
85 patients, laminoplasty in 44 patients, and laminectomy with posterior instrumented
fusion in 29 patients. Out of 158 patients, those with OPLL and CSM were 45 and 113,
respectively. In all, 10 patients developed C5 nerve root palsy postoperatively, out
of which 2 patients underwent laminoplasty, 6 patients had laminectomy, and 2 patients
had laminectomy with instrumented fusion. Palsy was higher in those with OPLL (8.3
vs. 1.25% in CSM), which was statistically significant.
Radiological: Mean preoperative cervical lordosis in palsy and non-palsy groups was
17.4 degree and 7.5 degree, respectively. Mean change in cervical lordosis following
ACDF and ACCF was 8.4 degree and 11.5 degree, respectively. Following laminectomy,
laminoplasty, and laminectomy with fusion, the change in cervical lordosis was 0.7
degree, 4.2 degree, and 21.3 degree, respectively. Mean C45 intervertebral angle preoperatively
in palsy and non-palsy group was 9.4 degree and 2.5 degree, respectively. Mean postoperative
C45 intervertebral angle in palsy and non-palsy group was 11.5 degree and 5.9 degree,
respectively. The mean change in the cervical lordosis and intervertebral angle between
palsy group and non-palsy group was not significant. The preoperative and postoperative
radiological measurements in both the groups are shown in [Table 3].
Table 3
Preoperative and postoperative radiological changes in both palsy and non-palsy group
|
Palsy group
|
Non-palsy group
|
|
Pre-op value (degrees)
|
Post-op value (degrees)
|
Pre-op value (degrees)
|
Post-op
value (degrees)
|
|
Cervical lordosis
|
|
ACDF
|
–
|
–
|
12.1
|
20.5
|
|
ACCF
|
–
|
–
|
8.1
|
19.7
|
|
Laminoplasty
|
22
|
22.5
|
22.3
|
26.5
|
|
Laminectomy
|
18.3
|
21
|
28.7
|
29.4
|
|
Laminectomy with fixation
|
10
|
18
|
4.3
|
25.6
|
|
C45 intervertebral angle
|
|
ACDF
|
–
|
–
|
2.15
|
5.65
|
|
ACCF
|
–
|
–
|
2.43
|
4.95
|
|
Laminoplasty
|
12.5
|
12.5
|
2.9
|
6
|
|
Laminectomy
|
10.5
|
12.5
|
5.45
|
7.24
|
|
Laminectomy with fixation
|
3
|
7.5
|
−0.5
|
5
|
Abbreviations: ACCF, anterior cervical corpectomy and fusion; ACDF, anterior cervical
discectomy and fusion.
Mean C45 foramen width ([Table 4]) in palsy group was 2.85 mm and in non-palsy group was 4.12 mm, which was statistically
significant (p < 0.01). Mean foramen width on the palsy side was significantly less as compared
with contralateral foramen width (2.5 vs. 3.29 mm). Side of the palsy correlated with
side of narrow foramen in 80% (8/10) of the patients. In 70% (7/10) of the patients
in palsy group, the foramen width on palsy side was < 3 mm.
Table 4
C45 foramen width as measured in CT scan at narrowest point in various groups
|
C45 foramen width (mean)
|
|
Palsy side
|
Contralateral side
|
|
2.5 mm
|
3.2 mm
|
|
Palsy group
|
Non-palsy group
|
|
2.85 mm
|
4.12 mm
|
Abbreviation: CT, computed tomography.
C5 palsy: Onset of palsy was seen after a mean postoperative period of 2.8 days and
mean MMT grade at time of onset was 2.9. All the patients were managed conservatively
with rest, physiotherapy, nonsteroidal anti-inflammatory drugs, and corticosteroids
in cases with severe palsy (MMT ≤ 2). All except one patient recovered to MMT ≥ 4
after a mean duration of 6.3 months, at final follow-up the mean MMT score was 4.1.
Discussion
Anterior and posterior decompression surgeries are relatively safe with good results
in myelopathy patients.[4] C5 palsy is one serious complication, which causes significant morbidity to patient
due to weakness of deltoid muscle with or without involvement of biceps.[7] In a recent meta-analysis,[6] the incidence varied from 0 to 26.4% in anterior surgeries and from 0 to 50% in
posterior surgeries. One cause for this wide variation is due to lack of a standard
definition; all patients with decrease in deltoid ± biceps power without involvement
of other muscles by one or more MMT grade were reviewed in our study. In our study
of 390 patients, 10 patients developed palsy involving C5 nerve root postoperatively.
Several causes have been suggested, first of which was inadvertent nerve injury by
Satomi et al.[4] But with improvements in surgical technique and technology, there was no significant
reduction in C5 palsy incidence.[8] In a study by Tanaka et al,[9] no abnormal events were observed during surgery with intraoperative neuromonitoring.
Hence, this theory no longer holds true. Tanaka N et al[9] and Hojo et al[10] proposed that increased cervical curvature was a risk factor of C5 palsy in posterior
surgery and Wang et al[11] proposed the same in anterior surgeries. In our study, mean change in lordosis for
anterior procedures was 9.95 degree (in ACDF, 8.4 degree and in ACCF, 11.5 degree)
which was significantly more than the preoperative values, but none of the patients
developed C5 palsy. Mean increase in curvature following posterior procedures was
8.7 degree, which was not statistically significant when compared with preoperative
values.
Local lordotic angle (C45 intervertebral angle) was proposed to cause foraminal stenosis,
leading to C5 palsy.[12]
[13] The mean change in intervertebral angle following anterior and posterior surgeries
was 7.07 and 20.78 degree, respectively. The change was significant in anterior and
laminectomy with fusion group, but not in laminectomy or laminoplasty group. Though
there was a 6.9% incidence of palsy in laminectomy with fusion group, it was 0% in
anterior procedures; hence a change in C45 intervertebral lordotic angle in itself
would not be a significant risk factor.
Posterior shift of the cord following decompression surgeries is the most common proposed
theory for C5 palsy. Posterior shift in the cord is seen after both anterior[14] and posterior surgeries.[15]
[16] Shift of the cord leads to anchoring of nerve root against the superior articular
facet,[17] and C5 nerve root being short, located at apex of lordosis and horizontal course
makes it more prone for palsy compared with other nerve roots.[17]
[18] Laminectomy involves removal of the entire posterior arch, thus allowing more posterior
shift of cord. Radcliff et al noted a higher incidence of C5 palsy in patients who
underwent laminectomy as compared with laminoplasty.[19] In our study, the incidence of palsy in laminectomy group was 7% and in laminoplasty
group was 4.5%. This was statistically significant (p < 0.01). In patients who undergo a posterior decompression, those with OPLL may have
a more posterior shift of cord compared with those without OPLL. Here, those with
OPLL and posterior surgery had a higher incidence of palsy compared with those without
OPLL (13 vs. 3.5%), which was statistically significant (p = 0.04). No measurement for the posterior cord shift was done in our study.
Imagama et al[20] reviewed 1,858 patients who underwent laminoplasty and found that C4–5 foramen is
significantly smaller in the palsy group. In our study, C4–5 foramen width on palsy
side was compared with contralateral side and non-palsy group. Mean width of the foramen
was significantly less on the palsy side when compared with the non-palsy side (2.5
vs. 3.2 mm) and the side of narrow foramen corresponded to the side of palsy in 80%
of the patients. Also, the mean foramen width in palsy group was significantly less
compared with non-palsy group (2.85 vs. 4.12 mm). On this basis, preoperative identification
of susceptible patients may be done, and a prophylactic foraminotomy would prevent
a serious C5 palsy. None of the patient in ours study underwent concomitant foraminotomy.
Sasai et al[21] and Yanase et al[22] reported that preoperative electromyography's could detect a pre-existing radiculopathy
and a concomitant foraminotomy might prevent palsy. Komagata et al[23] in a review of 305 patients reported a significant reduction in C5 palsy in the
group of patients who underwent prophylactic foraminotomy.
Hosono et al[24] proposed that the heat generated intraoperatively by high-speed drill caused C5
palsy as the temperatures rose to 174°C with diamond burr and to 77°C with steel burr.
They reported a higher incidence of C5 palsy in the diamond burr group. Takenaka et
al,[25] who noted a significant (82%) decrease in the incidence of C5 palsy by intraoperative
irrigation with cold saline while drilling, strengthened this theory. But, the decrease
in incidence was only in the early onset group, hence mechanisms other than heat were
supposed to cause C5 palsy. All the surgeries in our study were performed using steel
burr with saline at room temperature.
On the basis of the appearance of high intensity areas (HIA) on postoperative magnetic
resonance imaging (MRI) in patients with C5 palsy, Chiba et al[26] proposed that local reperfusion of the cord could cause C5 palsy; furthermore, the
level of HIA corresponded to the level of palsy. Few authors also reported the appearance
of HIA on postoperative MRI of patients with palsy,[27]; however, this theory could not explain the unilateral palsy in majority of patients.
Out of 44 patients who underwent laminoplasty, 2 developed C5 palsy. Palsy was seen
on the opening side in both the patients, but the difference was not significant statistically.
In our study, all patients underwent only open-door laminoplasty, and thus no conclusion
could be drawn regarding the technique of palsy.
Other risk factors proposed were old age,[14]
[28] male gender,[19] severity of myelopathy (or JOA score),[8] and number of decompression levels.[29] The mean age in palsy group and non-palsy group was 55.9 62.3 years, mean preoperative
JOA score of patients in palsy group and non-palsy group was 11.4 10.4, male to female
ratio in palsy and non-palsy groups was 1:1 and 1.2:1, respectively; the differences
in neither was statistically significant. Bose et al[30] concluded that intraoperative neuromonitoring with simultaneous sphincter electromyography
and transcranial motor evoked potentials can be effective in preventing C5 palsy,
but the studies by Fan et al[31] and Inoue et al[32] showed no proof that incidence of C5 palsy could be minimized with neuromonitoring.
Satisfactory recovery was achieved with conservative treatment in majority (90%) of
the patients in our study; no revision surgery was performed. One patient had persistent
palsy at final follow-up. Limitations of our study are less number in the palsy group,
no direct measurement of posterior shift of cord, and retrospective nature of study.
Conclusion
Age, preoperative JOA score, change in cervical lordosis, and C4–5 intervertebral
angle does not appear to be associated with C5 palsy. Posterior surgeries (as compared
with anterior surgeries), laminectomy (as compared with laminoplasty), C45 foraminal
stenosis, and OPLL were seen to be risk factors for C5 nerve root palsy. No one factor
explains the palsy in all cases, and hence the cause is multifactorial. Prophylactic
foraminotomy in patients with narrow C45 foramen may prevent this severe complication.
Majority of patients recovered with conservative management.
