Keywords
cervical spondylotic myelopathy - functional recovery - kyphotic - lordotic - nonlordotic
- outcomes - preoperative - sagittal alignment
Introduction
The role of cervical sagittal (CS) parameters in cervical surgeries is yet to be completely
understood. In the last decade, its influence on surgery for cervical spondylotic
myelopathy (CSM) has come to the forefront and CS balance has been identified as an
important determinant of radiological and clinical outcomes. Studies suggest that
CS balance closely relates to patients' health-related quality-of-life (HRQOL) scores
just as in the thoracolumbar spine.[1]
[2]
As far as impact of cervical parameters in CSM is concerned, it has been established
that sagittal imbalance (higher CS vertical axis [cSVA]) after surgery portends toward
poorer patient-reported outcomes.[3]
[4] Also, higher preoperative cSVA correlates with higher myelopathy disability.[4]
[5]
[6] Noticeably, the approach used,[7] type of surgery,[8] or the number of levels operated on[9] have not been shown to independently predict surgical outcomes. Several determinants
of postoperative neurological recovery such as preoperative myelopathy severity, duration
of symptoms, magnetic resonance imaging (MRI), and T2-weighted (T2W) signal intensities
(SI) are also well established in the literature.[9]
[10]
[11]
[12] However, evidence on the impact of preoperative sagittal curvature on the neurological
recovery is still evolving and needs further probing. This study aims to establish
the influence of preoperative curvature of the cervical spine (lordotic or nonlordotic)
on the neurological recovery and disability of the operated CSM patients at a minimum
1-year follow-up.
Materials and Methods
The study was proceeded after ethical approval of the institutional review board (EC/2/19/2008)
for using patient data and informed consent was taken from all the patients before
surgical intervention. We performed a retrospective analysis of data of consecutively
operated cases of degenerative cervical myelopathy (DCM) from March 2019 to April
2021 at our institute. The inclusion criteria were cases with clinical signs of myelopathy
with concordant spinal MRI changes, having age 18 or above with complete set of records,
and cervical lateral radiographs having properly visible endplates up to the C7 vertebra.
Cases excluded were ossification of the posterior longitudinal ligament (OPLL), tandem
stenosis, any previous spine surgery, inflammatory diseases (rheumatoid arthritis
or ankylosing spondylitis), infection/neoplasm, trauma, or those who had an incomplete
follow-up. [Fig. 1] depicts the flowchart of the methodology of the study.
Fig. 1 Flowchart depicting the methodology of the study (CSM, cervical spondylotic myelopathy;
DCM, degenerative cervical myelopathy; OPLL, ossification of the posterior longitudinal
ligament).
Radiographic Assessment
Measurements were done by standard lateral radiographs of cervical spine taken in
the neutral position with the upper extremities positioned at the side of the body
while maintaining a horizontal gaze. Sagittal parameters measured were (1) C2–C7 lordosis
(CL; in degrees); (2) C2–C7 sagittal vertical axis (cSVA; in mm); and (3) C7 slope
(C7S; in degrees; [Fig. 2]). Patient demographics were recorded and measurements were done with Surgimap V
2.3.2 (Nemaris Inc., New York City, NY) application by two spine surgeons on two separate
occasions and the mean of measurement was further analyzed. These assessors were blinded
to the outcomes of the patients. Subjects were grouped into two categories on the
basis of the preoperative cervical alignment ([Fig. 3]). Lordotic curvature (with Cobb angle >10 degrees) and nonlordotic curvature (including
neutral [Cobb angle 0–10 degrees] and kyphotic [Cobb angle < 0 degrees]).[13]
Fig. 2 Sagittal parameters measured were the following: (1) C2–C7 lordosis (CL, in degrees):
The Cobb angle between the lower endplates of C2 and C7 vertebral body. “ + ” denotes
lordosis, while “–” denotes kyphotic alignment. (2) C2–C7 sagittal vertical axis (cSVA,
in mm): The distance from the posterosuperior corner of C7 to a vertical line from
the center of the C2 vertebra. (Anterior shift of the plumbline was assigned “ + ”,
while the posterior shift was assigned “–”). (3) C7 Slope (C7S, in degrees): The angle
between the upper endplate of the C7 vertebral body and the horizontal.
Fig. 3 Cases were divided on the basis of preoperative alignment. (A) Lordotic and nonlordotic curvature. (B) Kyphotic alignment. (C) Neutral alignment.
To prevent heterogeneity of patient demographics, it was ensured that both groups
were comparable for the number of levels addressed, type of approach used, and preoperative
T2W MRI changes ([Table 1]). Qualitative MRI changes on T2W sequence included the presence of hyperintensity
signal in the cord. Quantitative changes in T2W sagittal MRI were assessed using signal
intensity ratio (SI Ratio). It was defined as the ratio between the intensity at the
area of the greatest cord SI change and the cerebrospinal fluid (CSF) behind the spinal
cord at C2, calculated with 100-pixel circles at both places. The analysis was done
using ImageJ software (National Institutes of Health, Bethesda, MD)[5] by two spine surgeons who were blinded to the functional outcomes.
Table 1
Demographic comparison between lordotic and nonlordotic groups
|
Parameters
|
Cobb > 10 (N = 78)
Lordotic
|
Cobb ≤ 10 (N = 46)
Kyphotic/neutral
|
p-Value
|
|
Sex (M:F)
|
60:10
|
39:2
|
0.1
|
|
Age (y)
|
54.1 ± 13.08
|
51.4 ± 13.37
|
0.3
|
|
ASAPS grade (1:2:3)
|
41:23:6
|
27:10:4
|
0.6
|
|
Pre-op severity (mJOA)
|
11.7 ± 2.18
|
11.5 ± 2.06
|
0.6
|
|
No. of levels addressed (mean)
|
2.15 ± 0.93
|
2.15 ± 0.98
|
0.9
|
|
Approach used (anterior:posterior)
|
45:32
|
30:15
|
0.5
|
|
Duration (mo)
|
5.6 ± 4.4
|
6.0 ± 4.9
|
0.6
|
|
MRI SI ratio
|
0.64 ± 0.1
|
0.66 ± 0.1
|
0.3
|
|
Presence of MRI hyperintensity
|
88.6%
|
85.4%
|
0.7
|
|
Mean follow-up (y)
|
2.5 ± 1.44
|
2.49 ± 1.51
|
0.9
|
Abbreviations: ASAPS, American Society of Anesthesiologists Physical Status; mJOA,
modified Japanese Orthopaedic Association; MRI, magnetic resonance imaging; SI, signal
intensity.
Surgical Procedure
All the patients underwent cervical decompression surgeries, performed by three senior
spine surgeons of the ortho-spine department of the institute. The choice of anterior
or posterior surgery was at the discretion of the surgeon. Factors such as number
of levels involved, cervical alignment, presence or absence of retrovertebral compression,
and patient's surgical capacity were weighed in decision-making. Anterior surgeries
included anterior cervical diskectomy and fusion (ACDF), anterior cervical corpectomy
and fusion (ACCF), or hybrid surgery, whereas posterior surgeries included cervical
laminectomy with or without lateral mass fixation and laminoplasty. In few cases,
combined anterior and posterior surgery was done.
Outcome Measurement
Both lordotic curvature (with Cobb angle > 10 degrees) and nonlordotic curvature (including
neutral [Cobb angle 0–10 degrees] and kyphotic [Cobb angle < 0 degrees]) were analyzed
for various baseline characteristics and checked for comparability. Cases were followed
at baseline, postoperatively 1 year, and at last follow-up. Functional outcomes were
compared using the change in modified Japanese Orthopaedic Association (mJOA) scale
and Nurick grade. The functional recovery rate for the mJOA scale (mJOArr) was calculated
using Hirabayashi's method[14]:
Nurick grade recovery rate (Nurick RR)[14] was calculated as follows:
Correlations between the CS parameters and outcomes scores were calculated.
Statistical Analysis
Statistical analysis was performed using IBM SPSS version 23.0 (IBM Corp., Armonk,
NY, United States). Continuous variables were described using means, standard deviations,
and ranges. Quantitative data were first tested for its normality and homogeneity
of variance and according to different situations. We used the Shapiro–Wilk test of
normality for the continuous variable for linear regression. Categorical variables
were summarized using frequencies and percentages. They were tested by Student's t-test or Mann–Whitney U test, while qualitative data were tested by chi-square test to compare the differences
between the lordotic and nonlordotic groups. Spearman's correlational coefficients
were calculated to assess associations between CS parameters and outcome measures.
The interobserver reliability and intraobserver reproducibility were assessed using
intraclass correlation coefficient (ICC) at 95% confidence interval. The level of
significance for all analyses was defined as α equal to 0.05.
Results
In our analysis of 124 cases, 63.1% (78 cases) were lordotic and 36.9% (46 cases)
were nonlordotic preoperatively. In the nonlordotic group, 24.6% (32 cases) had neutral
alignment, while 12.3% (14 cases) had kyphotic alignment. The mean Cobb angle of the
lordotic group was 23.57 ± 9.1 degrees (11–50 degrees), whereas the mean Cobb angle
of the nonlordotic group was 0.89 ± 6.5 degrees (–11 to 10 degrees). Demographic characteristics
of the two groups were not significantly different from each other and the groups
were comparable in baseline demographics ([Table 1]). There was significant increase in mean mJOA score (∆mJOA = 3.2 ± 0.46; p < 0.001) and reduction in Nurick grade (∆Nurick = 1.4 ± 0.32; p < 0.001) after surgical intervention. Overall change in the mean sagittal parameters
at the last follow-up are summarized in [Table 2]. The mean change in mJOA, Nurick grade, and functional recovery rate were not significantly
different between the lordotic and nonlordotic groups ([Table 3]). The surgical approach used in the lordotic group was 57.1% anterior, 41.4% posterior,
and 1.4% a combined anterior and posterior approach, whereas in the nonlordotic group,
65.9% were anterior, 31.7% posterior, and 2.4% a combined anterior and posterior surgery.
In the nonlordotic group, patients who received anterior surgery had a significantly
better mJOArr (60.4 ± 30.2 vs. 41.9 ± 37.1%; p = 0.04) than those who were approached posteriorly. Lordotic cases had a similar
improvement with either approach. Overall, the mean number of levels operated in the
anterior surgery was 1.8 ± 0.81, whereas that in the posterior surgery was 2.80 ± 0.82
(p < 0.001).
Table 2
Overall change in parameters at the final follow-up
|
Parameters
|
Preoperative
|
Last follow-up
|
p value
|
ICC
|
|
cSVA
|
18.5 ± 12.73
|
21.1 ± 12.59
|
0.001
|
0.80
|
|
C2–C7 Cobb
|
15.1 ± 13.73
|
9.9 ± 10.53
|
< 0.001
|
0.89
|
|
C7 slope
|
25.1 ± 9.71
|
19.9 ± 8.15
|
0.089
|
0.80
|
|
mJOA scale
|
11.68 ± 2.12
|
14.86 ± 2.58
|
< 0.001
|
–
|
|
Nurick grade
|
3.1 ± 0.97
|
1.7 ± 1.29
|
< 0.001
|
–
|
Abbreviations: cSVA, cervical sagittal vertical axis; ICC, intraclass correlation
coefficient; mJOA, modified Japanese Orthopaedic Association.
Table 3
Comparison of lordosis versus nonlordosis group at the final follow-up
|
N = 124
|
Cobb > 10 (N = 78)
Lordotic
|
Cobb ≤ 10 (N = 46)
(kyphotic/neutral)
|
p-Value
|
|
Fraction
|
63.1%
|
36.9%
Kyphotic (n = 14;12.3%)
Neutral (n = 32; 24.6%)
|
0.1
|
|
Mean Cobb (degrees)
|
23.57 ± 9.1
|
0.89 ± 6.5
|
0.03
|
|
∆Cobb (mean ± SD; degrees)
|
–10.25 ± 5.6
|
+4.01 ± 2.9
|
0.006
|
|
% of anterior surgery
|
57.1
|
67.8
|
0.6
|
|
∆mJOA
|
3.21 ± 1.68
|
3.17 ± 2.37
|
0.9
|
|
∆Nurick
|
1.50 ± 0.93
|
1.32 ± 0.92
|
0.5
|
|
Nurick RR (%)
|
49.19 ± 30.11
|
48.57 ± 33.89
|
0.6
|
|
mJOArr (%)
|
52.5 ± 25.84
|
54.56 ± 33.26
|
0.9
|
|
Final alignment
|
Lordotic
51.12%
|
Lordotic
20.5%
|
–
|
|
Neutral
37.14%
|
Neutral
62.4%
|
|
Kyphotic
5.7%
|
Kyphotic
17.1%
|
Abbreviations: mJOA, modified Japanese Orthopaedic Association; mJOArr, functional
recovery rate for the mJOA scale; RR, recovery rate; SD, standard deviation.
The lordotic group had a significant loss of mean lordosis (–9.25 ± 1.38 degrees;
p < 0.001), whereas the nonlordotic group gained lordosis significantly (+4.01 ± 1.39 degrees;
p < 0.003) at the final follow-up. In the nonlordotic group, patients who gained lordosis
(78.1%, n = 36) from their baseline alignment had better recovery rates than those who lost
lordosis (21.9%, n = 10). This difference in recovery rates, however, did not meet the statistical significance
(mJOArr, p = 0.08, and Nurick RR, p = 0.1; [Table 4]). Statistically significant correlations between sagittal parameters and the functional
scores observed in the two groups are listed in [Table 5].
Table 4
Restoration of lordosis in nonlordotic group and functional outcome
|
Nonlordotic group
|
Percentage
|
Mean ∆Cobb (degrees)
|
mJOArr
|
Nurick RR
|
|
Gain in lordosis ∆Cobb (positive)
|
78.04
|
6.8
|
57.07%
|
53.6%
|
|
Loss in lordosis ∆Cobb (negative)
|
21.96
|
–6.5
|
42.4%
|
37.61%
|
|
p-Value
|
–
|
–
|
0.08
|
0.1
|
Abbreviations: mJOA, modified Japanese Orthopaedic Association; mJOArr, functional
recovery rate for the mJOA scale; RR, recovery rate.
Table 5
Correlations between sagittal parameters and functional scores. Increasing sagittal
imbalance (higher cSVA) correlates with higher preoperative disability (lower mJOA
score) in nonlordotic spine only, while high cSVA postoperatively correlates with
poorer functional recovery rates (mJOArr) in both lordotic and nonlordotic groups
|
Correlations
|
Nonlordotic
|
Lordotic
|
|
Preoperative cSVA and preop mJOA (disability)
|
r = –0.336, p = 0.01
|
r = 0.4, p = 0.1
|
|
Postoperative cSVA and mJOArr
|
r = –0.342, p = 0.02
|
r = –0.428, p < 0.001
|
Abbreviations: cSVA, cervical sagittal vertical axis; mJOA, modified Japanese Orthopaedic
Association; mJOArr, functional recovery rate for the mJOA scale.
Discussion
The importance of CS parameters is now increasingly being recognized in planning and
assessing the results of patients undergoing surgery for CSM. Conventionally, preoperative
cervical alignment dictates the approach to be used for surgery. Posterior decompressive
surgeries are preferred in patients with lordotic alignment relying on the cord fallback
posteriorly, whereas anterior surgeries are advocated in nonlordotic alignment to
restore lordosis.[15] However, the literature seems to be divided with regard to the influence of preoperative
alignment on the functional outcome after surgery. Shamji et al[16] reported that patients with preoperative kyphotic alignment showed inferior neurological
improvement than those with lordotic curvature. On the other hand, there are studies
that show that both lordotic and nonlordotic alignment had comparable final functional
results.[13]
[17] In this study, we set to determine the influence preoperative alignment had on postoperative
functional outcome in 124 cases of pure CSM.
In our analysis, patients with preoperative nonlordotic alignment were not associated
with inferior functional results when compared with cases with lordotic alignment.
In the nonlordotic cohort, anterior surgery had better functional recovery rates (p = 0.04) than those who had posterior surgery. Also, this group had gained mean lordosis
of +4.01 degrees at the final follow-up. In all, 78.04% of patients who had gained
lordosis showed higher mean recovery rates than 21.6% patients who had lost lordosis
(mJOA: 57.07 vs. 42.4%) at the last follow-up. This difference, however, could not
reach statistical significance (p = 0.08) probably due to a lesser number of kyphotic cases (14 cases, 12.3% overall)
in the nonlordotic cohort ([Table 3]). The authors suggest that realignment toward lordosis plays an important role in
the cases which are nonlordotic at baseline. The hypothesis given by Batzdorf and
Batzdorf[18] states that gain of lordosis allows spinal cord to fall back from the anteriorly
draping compressive forces of the disk osteophytes complex present in a kyphotic spine.
Also, both global and segmental kyphoses of cervical spine alter the biomechanics
and hasten the degenerative process leading to early arthritic changes with predisposition
to adjacent segment degeneration.[19] Direct decompression and improvement in lordosis using anterior approach allows
an environment of neurological recovery in such cases as demonstrated in our analysis
with the anterior approach surgery faring better than the posterior surgery in the
nonlordotic group.
Our results echo with the study of Kaptain et al,[13] who also divided their study cohort on basis of preoperative sagittal alignment
into lordotic, straight, and kyphotic curvatures comparing the clinical outcomes of
each. They reported that neither preoperative nor postoperative alignment influenced
the functional outcome. They went on to suggest that efforts to prevent kyphotic deformity
may rather be futile. Similarly, Jain et al[17] showed that the presence of segmental kyphosis preoperatively does not lead to inferior
functional results when compared with cases having preoperative lordosis after decompressive
surgery. Our results, however, contradict the study by Shamji et al,[16] who reported that patients with preoperative kyphotic alignment were associated
with inferior neurological recovery than those who had lordotic alignment at the baseline.
They assessed the outcomes by seeing improvements in the mJOA score, Nurick grade,
and the 30-second walk test. Noticeably, as compared with this study, they had a higher
proportion of kyphotic cases present in their cohort (34 vs. 12%). Also, despite achieving
significant correction in sagittal alignment postoperatively in their kyphotic group
(mean ∆Cobb of +13 ± 6 degrees), they still reported poorer myelopathy recovery as
compared with the lordotic group.
The majority of surgical patients with CSM exhibit some neurological improvement following
intervention as found in our analysis. Overall, there was significant improvement
in the mJOA score (p < 0.001) and Nurick grade (p < 0.001) at the final follow-up. Preoperative myelopathy severity (preoperative mJOA)
correlated with the preoperative cSVA (p = 0.01; r = –0.336) in the nonlordotic group only and no such correlation was seen in the lordotic
alignment group ([Table 5]). Similar findings have been reported by previous authors in which increasing cSVA
correlated with a higher preoperative disability in the kyphotic (nonlordotic) patients
and not in the lordotic spine.[5]
[6] We observed postoperatively the increasing cSVA (sagittal imbalance), that is, a
shift of alignment from lordotic to neutral and then finally kyphosis, negatively
correlated with the functional outcome, irrespective of whether the preoperative alignment
is lordotic or nonlordotic ([Table 5]; [Figs. 4] and [5]). This effect could be explained with increased anterior cord distortion with increasing
sagittal imbalance as the neck alignment goes from lordosis to kyphosis. These findings
emphasize improving the sagittal alignment as a potential surgical objective in cases
of CSM.
Fig. 4 Postoperative C2–C7 sagittal vertical axis (cSVA) showed statistically significant
negative correlation with functional recovery rate in posterior surgeries. Radiographs
showing good functional recovery with smaller cSVA as shown in (A) a case of laminoplasty than having a larger cSVA as shown in (B) a case of laminoplasty and (C) a case of posterior laminectomy with lateral mass fixation.
Fig. 5 Postoperative C2–C7 sagittal vertical axis (SVA) showed statistically significant
negative correlation with functional recovery rate in anterior surgeries as well.
Radiographs showing good functional recovery with smaller cSVA as seen in (A) a case of anterior cervical discectomy and fusion than having larger cSVA as shown
in (B) a case of postoperative anterior cervical corpectomy with mesh cage and plating.
Our analysis shows that the functional results of lordotic and nonlordotic cervical
spine were comparable after surgery for CSM provided there was some gain in lordosis
in the cases with baseline nonlordotic alignment. The anterior approach faring better
than the posterior approach further stresses the importance of improvement in alignment
in surgery. However, the magnitude of increase in lordosis, which can be considered
sufficient, is a matter of further analysis. Previous reports suggest that the amount
of lordotic correction (small or large) does not seem to affect the functional outcome
if the final alignment is lordotic.[20] Also, increasing lordosis in preoperatively lordotic spines does not seem to give
added functional advantage.[21] At the same time, enormous increase in lordotic alignment may invite complications
such as C5 palsy.[22] Thus, it seems defining the optimum correction of alignment and the appraisal of
its impact on functional outcome warrant a more robust analysis with larger randomized
studies.
We acknowledge various limitations of the study. First, the C0–C2 segment lordosis
was not taken into consideration, which has significant contribution to the overall
lordosis of the cervical spine. The criterion assigned for neutral curvature (Cobb
angle: 0–10 degrees) might constitute to analytical bias. Other limitation would be
not analyzing the global alignment parameters including the spinopelvic parameters
in a study discussing the impact of the regional cervical parameters on the functional
outcome. Also, the retrospective nature of the analysis and a smaller subset of kyphotic
patients in the cohort (12.3%) might have reduced the power of the study. We ensured
blinding wherever possible to reduce biases in assessment.
Conclusion
Decompression and stabilization seems to be the mainstay of the surgical treatment
for CSM. We report the noninferiority of the functional outcome in the cases with
preoperative nonlordotic alignment as compared with lordotic alignment of the cervical
spine. Sagittal imbalance (high cSVA) postoperatively correlates with poorer functional
recovery rates. Also, increasing sagittal imbalance correlates with a higher preoperative
disability in the nonlordotic spine, a finding not seen in the lordotic spine. Still
realignment towards lordosis in cases with baseline nonlordotic spines have a biomechanical
advantage and may lead to improved results. We recommend larger randomized studies
to further analyze the true impact of sagittal alignment on outcomes. As a goal for
decompressive surgery for CSM, consideration should be given to restoration of cervical
lordosis and correction of CS imbalance when present.
