Abstract
Study Design:
Retrospective multicenter study.
Objective:
Posterior decompression surgery for cervical spondylotic myelopathy (CSM) and ossification of the posterior longitudinal ligament (OPLL) is a common surgery; however, it can cause postoperative cervical deformity (CD). The purpose of this study was to investigate the risk factors for CD.
Methods:
The participants were 193 patients underwent laminoplasty or laminectomy for CSM or OPLL. CD was defined as a C2-7 sagittal vertical axis (SVA) ≥ 40 mm or a cervical lordosis angle (CL) ≤ −10°. The participants were divided into 2 groups: NCD (without CD before surgery), CD (with CD before surgery). NCD group was divided based on the presence of CD 1 year after surgery as follows: postoperative CD (PCD) and no PCD (NPCD).
Results:
There were 153 patients (NCD), 40 (CD), 126 (NPCD), and 27 (PCD). There was significant difference in the number of decompressed lamina (NPCD: PCD = 4.1:4.5), the presence of C2 decompression (2: 11%), and C5 palsy (0: 11%). The risk factors for onset of CD, PCD, and CL ≤ −10° as assessed by multiple logistic regression analysis were preoperative C2-7 SVA ≥ 30 mm (odds ratio [OR]: 19.0), decompression of C2 or C7 lamina (OR 3.1), and preoperative CL ≤ 2° (OR 42.0), respectively.
Conclusions:
To prevent postoperative CD, it is important to avoid decompression of the C2 or C7 lamina. Moreover, in case with C2-7 SVA ≥ 30 mm or CL ≤ 2° before surgery, it is important to explain the risks and consider adding fusion surgery.
Keywords: cervical deformity, cervical malalignment, cervical posterior decompression surgery, laminoplasty, cervical kyphosis
Introduction
Posterior cervical decompression surgery, which comprises laminectomy and laminoplasty, is an accepted treatment for patients with progressive myelopathy caused by cervical spondylotic myelopathy (CSM) or ossification of the posterior longitudinal ligament (OPLL). However, laminectomy can lead to instability, kyphotic deformity, formation of a post-laminectomy membrane, and late neurological deterioration.1-6 For instance, it has been reported that after laminectomy, cervical kyphosis deteriorates in 6 to 30% of cases.7-9 Therefore, various modalities of laminoplasty have been adopted to prevent these complications.10,11 However, even with laminoplasty, the loss of cervical lordosis after surgery is a common complication and remains an important issue.12-14
There are many reports stating that the most common cervical deformity is kyphosis, which leads to neurological symptoms including neck pain and myelopathy.15-17 Moreover, it has been suggested that alterations of the C2-7 sagittal vertical axis (SVA) lead to poor health-related quality of life.18-20 On the other hand, Suda et al reported that the risk of poor recovery rate for Japanese Orthopaedic Association score was local kyphosis exceeding 13°. 21 Recently, Smith et al evaluated the adult cervical spine deformity (ACSD) using cervical kyphosis, C2-7 SVA, and chin-brow angle. 22 They indicated that symptomatic ACSD has a negative impact on all EQ-5D domains, and that the mean EQ-5D of ACSD patients is comparable to the 25th percentile values for blindness/low vision, emphysema, renal failure, and stroke. Therefore, it is important to prevent cervical deformity (CD) after posterior cervical decompression surgery. However, whether cervical spine alignment changes occur when patients with CD undergo posterior cervical decompression surgery is unclear, and the risk factors for postoperative cervical deformity (PCD) are also unknown.
The purpose of the present study was to investigate the risk factors for PCD in patients who underwent posterior cervical decompression surgery for CSM and OPLL, and the changes in spine alignment in patients with preoperative ACSD.
Methods
Institutional Review Board Approval
This study protocol was approved by the Institutional Review Board (IRB) of our university hospital. The IRB approval number is 19-242. The consent from the patients was taken in an opt-out format.
Subjects
A total of 287 patients who underwent posterior decompression surgery (laminectomy, double-door laminoplasty, or open-door laminoplasty except for fenestration) for CSM and OPLL at our hospital or at satellite hospitals (9 hospitals) from April 2010 to October 2019 were enrolled in this study. The exclusion criteria were as follows: (1) unavailable standing radiographs, (2) less than 1-year follow-up after surgery, and (3) unclear radiographic parameters.
CD was defined as the presence of at least one of the following: cervical lordosis (CL; the angle between C2 endplate and C7 lower endplate) < −10° or C2-7 SVA > 4 cm. 22 The definition of chin brow vertical angle >25° was excluded because the face was often not included in the image range of the whole spine radiographs. The subjects were divided as follows: (1) no CD (NCD; patients without CD before surgery) group, (2) CD (patients with CD before surgery) group. Moreover, NCD group was divided based on the presence of CD as follows: (3) PCD group and (4) no PCD (NPCD) group.
Data Collection and Radiographic Parameters
The data collected was age, sex, height, weight, body mass index (BMI), smoking history, alcohol consumption, surgical procedure (laminectomy, double-door laminoplasty, or open-door laminoplasty), number of decompressed vertebrae, instability (defined as translation over 3 mm using sitting X-ray), presence of C2 or C7 decompression, estimated blood loss, length of operation, hospital stay (days), and modified Japanese Orthopaedic Association (mJOA) score.
The radiographic parameters assessed by standing X-ray were T1 slope (TS), CL, TS minus CL (TS-CL), C2-7 SVA (the distance between a plumb-line from the center of the C2 vertebra and the posterior superior corner of C7), and slope of McGregor’s line (MCGS; the angle between McGregor’s line and a horizontal line [positive indicates that the head posture rotated anteriorly]).
Statistical Analysis
Statistical analyses were performed using SPSS version 25 (IBM Corp., Armonk, NY, USA). P < 0.05 was considered statistically significant. Continuous variables are presented as mean ± SD. Categorical variables were evaluated using the chi-square test or Fisher’s exact test. A paired t-test and an independent t-test was used to assess the change in radiographic alignment after surgery and compare the data of each group. Receiver operating characteristic (ROC) analysis was used to identify the cut-off values of continuous variables. The risk factors of CD were identified using multiple logistic regression analysis.
Results
The Comparison of CD and NCD Group
Patients’ characteristics
Of the initial 287 patients, 82 were excluded due to a lack of standing X-ray photographs before or 1 year after surgery, 6 were excluded due to having been transferred to another hospital or department where they could not be followed up for at least 1 year, and 6 were excluded because the X-ray photographs were not clear and the parameters could not be evaluated (Figure 1). Finally, 193 patients were included. CD and NCD groups were 40 and 153 patients, respectively. Of the 153 in NCD group, 27 developed PCD 1 year after surgery, and the remaining 126 did not develop CD 1 year after surgery (NPCD group).
Figure 1.
Flowchart of patient recruitment. From the screened 287 patients, 94 were excluded: 82 due to the absence of standing X-ray photographs, 6 due to having been transferred to another hospital or department, and 6 due to unclear X-ray photographs. A total of 193 patients were finally enrolled. The cervical deformity (CD) group, the no CD (NCD) group, no iatrogenic CD (NICD) group, and the iatrogenic CD (ICD) group comprised 40 153 126, and 27 patients, respectively.
Table 1 shows the patients’ characteristics in NCD and CD groups. There were no significant differences in mean age (NCD 68.6 ± 10.6 and CD 67.5 ± 13.1, P = 0.556) and male proportion (NCD 106 [69.3%], CD 32 [80%], P = 0.181). Height and weight in the CD group (163.1 ± 10.2 cm and 65.8 ± 17.6 kg) was significantly higher than that in the NCD group (158.6 ± 9.5 cm and 59.2 ± 13.0 kg, P = 0.006 and 0.011). However, there was no significant difference in BMI between the both groups. Current or past smoking rates in the CD group (63%) were significantly higher than those in the NCD group (39%, P = 0.001).
Table 1.
The Characteristics of Patients in NCD and CD Group.
| Group | NCD | CD | |
|---|---|---|---|
| Number of patients | 153 | 40 | P |
| Age | 68.6 ± 10.6 | 67.5 ± 13.1 | 0.556 |
| Sex (Male) | 106 (69.3%) | 32 (80%) | 0.181 |
| Height (cm) | 158.4 ± 9.4 | 163.1 ± 10.2 | 0.006** |
| Weight (kg) | 59.2 ± 13.0 | 65.8 ± 17.6 | 0.011* |
| Body mass index | 23.4 ± 3.9 | 24.4 ± 4.6 | 0.204 |
| Smoking (Current or Previous, No history) No answer; 15 in NCD, 2 in CD | 59, 79 (39%, 52%) | 24, 14 (63%, 37%) | 0.001** |
| Drinking alcohol (Yes, No) No answer; 18 in NCD, 3 in CD | 46, 89 (30%, 58%) | 11, 26 (30%, 70%) | 0.752 |
| Pathology (CSM, OPLL) | 106, 47 (69.3%, 30.7%) | 28, 12 (70%, 30%) | 0.930 |
| Surgical procedure(Double door, Open door, Laminectomy) | 149, 3, 1 (97%, 2%, 1%) | 38, 1, 1 (95%, .5%, 2.5%) | |
| Number of decompressed lamina | 4.2 ± 0.8 | 4.3 ± 0.7 | 0.438 |
| Instability | 26 (17.0%) | 1 (2.5%) | 0.019* |
| C2 decompression (Yes) | 5 (3.3%) | 2 (5%) | 0.445 |
| C7 decompression (Yes) | 57 (37.3%) | 17 (42.5%) | 0.544 |
| Estimated blood loss (ml) | 86.4 ± 109.1 | 91.7 ± 83.3 | 0.778 |
| Operative time (minute) | 112.8 ± 38.7 | 127.4 ± 38.4 | 0.098 |
| C5 palsy | 3 (2.0) | 1 (2.5%) | 0.608 |
| Hospital stays (day) | 22.6 ± 14.5 | 21.2 ± 11.8 | 0.559 |
| Revision surgery for CD | 0 | 2 (5%) | 0.042* |
*P < 0.05, **; P < 0.01.
Abbreviations: NCD; no cervical deformity, CD; cervical deformity, CSM; cervical spondylotic myelopathy, OPLL; Ossification of posterior longitudinal ligament.
The breakdown of pathology were 106 patients with CSM and 47 patients with OPLL in the NCD group, and 28 patients with CSM and 12 patients with OPLL in the CD group (P = 0.930). Regarding the surgical procedure, almost all patients in this study underwent a double-door laminoplasty (149 of 153 cases [97.4%] in NCD and 38 of 40 cases [95%] in CD). Open door laminoplasty was performed in 3 patients in the NCD group and 1 patient in the CD group, and laminectomy was performed in only 1 patient in both groups. The rate of cervical instability (17%) in NCD group was significantly higher than that (2.5%) in CD group (P = 0.019). There were no significant differences in number of decompressed vertebrae, C2 decompression, C7 decompression, estimated blood loss, operative time, C5 palsy, or hospital stay in both groups. However, the incidence of revision surgery was significantly higher in CD group (5%, P = 0.042).
Change in radiographic parameters
Table 2 shows the change in radiographic parameters in NCD and CD group 1 year after surgery.
Table 2.
The Change of Radiographic Parameters in NCD and CD Group.
| NCD (153 case) | CD (40 case) | P | ||
|---|---|---|---|---|
| TS | Before surgery | 27.8 ± 8.7 | 36.5 ± 10.9 | 0.000*** |
| 1 year after | 26.5 ± 11.2 | 31.8 ± 12.3 | 0.009** | |
| P | 0.062 | 0.000*** | ||
| CL | Before surgery | 16.7 ± 12.2 | 11.2 ± 15.1 | 0.016* |
| 1 year after | 13.0 ± 14.6 | 4.2 ± 22.7 | 0.025* | |
| P | 0.000*** | 0.002** | ||
| TS-CL | Before surgery | 11.1 ± 9.3 | 25.4 ± 10.7 | 0.000*** |
| 1 year after | 13.5 ± 12.5 | 27.6 ± 17.5 | 0.000*** | |
| P | 0.008** | 0.328 | ||
| C 2-7 SVA | Before surgery | 21.3 ± 10.9 | 50.2 ± 8.9 | 0.000*** |
| 1 year after | 25.9 ± 13.4 | 50.2 ± 17.7 | 0.000*** | |
| P | 0.000*** | 0.995 | ||
| MCGS | Before surgery | −1.7 ± 9.2 | 5.4 ± 10.6 | 0.000*** |
| 1 year after | −2.1 ± 7.7 | 0.5 ± 12.2 | 0.240 | |
| P | 0.479 | 0.045* | ||
| mJOA score | Before surgery | 11.3 ± 2.5 | 10.8 ± 3.0 | 0.355 |
| 1 year after | 13.4 ± 2.6 | 12.6 ± 3.2 | 0.218 | |
| P | 0.000*** | 0.003** |
*P < 0.05, **; P < 0.01, ***;P < 0.0001.
Abbreviations: NCD; no cervical deformity, CD; cervical deformity; TS; T1 slope, CL; cervical lordosis, C2-7 SVA; C2-7 sagittal vertical axis, MCGS; slope of McGregor’s line, mJOA score; modified Japanese orthopedic association score.
Preoperative TS (36.5 ± 10.9°), TS-CL (25.4 ± 10.7°), C2-7 SVA (50.2 ± 8.9 mm), and MCGS (5.4 ± 10.6) in CD group was significantly higher than those (TS: 27.8 ± 8.7°, TS-CL: 11.1 ± 9.3°, C2-7 SVA: 21.3 ± 10.9 mm, and MCGS: −1.7 ± 9.2°) in NCD group. Preoperative CL (11.2 ± 15.1°) in CD group was significantly lower than that (16.7 ± 12.2°) in NCD group. Moreover, these significant differences remained even after the surgery.
CL was significantly reduced 1 year after surgery in both groups. In particular, CL in the CD group decreased by 7° (P = 0.002). On the other hand, C2-7 SVA in CD group did not change 1 year after surgery. The line of sight of the CD group shown in MCGS was significantly lower. However, 1 year after surgery, MCGS in the CD group had decreased significantly (the line of sight was raised, P = 0.045), and no significant difference was observed in both groups.
The mJOA score in NCD (11.3 ± 2.5 to 13.4 ± 2.6, P < 0.0001) and CD (10.8 ± 3.0 to 12.6 ± 3.2, P = 0.003) groups improved significantly 1 year after surgery. However, there were no significant differences in both groups before and 1 year after surgery.
The Comparison of NPCD and PCD Group
Patients’ characteristics
Table 3 presents the patients’ characteristics in NPCD and PCD groups. There were no significant differences in mean age (NPCD 68.5 ± 10.3 and PCD 69.4 ± 12.4, P = 0.699) and male proportion (NPCD 88 [69.8%], PCD 18 [66.7%], P = 0.746). There was also no significant difference in height, weight, BMI, smoking history, and drinking history in both groups.
Table 3.
The Characteristics of Patients in NPCD and PCD Group.
| Group | NPCD | PCD | |
|---|---|---|---|
| Number of patients | 126 | 27 | P |
| Age | 68.5 ± 10.3 | 69.4 ± 12.4 | 0.699 |
| Sex (Male) | 88 (69.8%) | 18 (66.7%) | 0.746 |
| Height (cm) | 158.6 ± 9.5 | 157.1 ± 9.2 | 0.444 |
| Weight (kg) | 59.4 ± 13.8 | 58.7 ± 8.7 | 0.808 |
| Body mass index | 23.4 ± 4.0 | 23.8 ± 3.4 | 0.557 |
| Smoking (Current or Previous, No history) No answer; 10 in NICD, 5 in ICD | 51, 65 (44%, 56%) | 8, 14 (36%, 64%) | 0.329 |
| Drinking alcohol (Yes, No) No answer; 13 in NICD, 5 in ICD | 40, 73 (35%, 65%) | 6, 16 (27%, 73%) | 0.366 |
| Pathology (CSM, OPLL) | 86, 40 (68.3%, 31.7%) | 19, 8 (70.4%, 29.6%) | 0.873 |
| Surgical procedure (Double door, Open door, Laminectomy) | 122, 3, 1 | 27, 0, 0 | |
| Number of decompressed lamina | 4.1 ± 0.7 | 4.5 ± 0.8 | 0.015* |
| Instability | 21 (16.7%) | 5 (18.5%) | 0.503 |
| C2 decompression (Yes) | 2 (1.6%) | 3 (11.1%) | 0.039* |
| C7 decompression (Yes) | 44 (34.9%) | 13 (48.1%) | 0.197 |
| Estimated blood loss (ml) | 80.9 ± 99.3 | 110.4 ± 144.3 | 0.319 |
| Operative time (minute) | 112.8 ± 38.7 | 127.6 ± 49.2 | 0.089 |
| C5 palsy | 0 | 3 (11.1%) | 0.005** |
| Hospital stays (day) | 22.0 ± 14.0 | 25.3 ± 16.6 | 0.286 |
| Revision surgery for CD | 0 | 0 |
*P < 0.05, **; P < 0.01.
Abbreviations: NPCD; no postoperative cervical deformity, PCD; postoperative cervical deformity, CSM; cervical spondylotic myelopathy, OPLL; Ossification of posterior longitudinal ligament.
The breakdown of pathology were 86 patients with CSM and 40 patients with OPLL in the NPCD group, and 19 patients with CSM and 8 patients with OPLL in the PCD group (P = 0.873). The ratio of the double-door laminoplasty were 122 of 126 cases (96.8%) in NPCD and 27 of 27 cases (100%) in PCD. Open door laminoplasty and laminectomy was performed in 3 patients and 1 patient in the NPCD group, respectively. There were no significant differences in spinal instability, C7 decompression, estimated blood loss, operative time, or hospital stay in both groups.
However, the number of decompressed vertebrae (4.5 ± 0.8) and C2 decompression (11.1%) in the PCD group was significantly higher than those (number of decompressed vertebrae: 4.1 ± 0.7, P = 0.015 and C2 decompression: 1.6%, P = 0.039) in the NCD group. Moreover, the incidence of C5 palsy was significantly higher in the PCD group (11.1%) than in the NPCD group (0%, P = 0.005).
Table 4 demonstrates the change in radiographic parameters in NPCD and PCD group 1 year after surgery.
Table 4.
The Change of Radiographic Parameters in NPCD and PCD Group.
| NPCD (126 case) | PCD (27 case) | P (ANOVA) | ||
|---|---|---|---|---|
| TS | Before surgery | 27.3 ± 8.6 | 29.9 ± 9.1 | 0.161 |
| 1 year after | 25.6 ± 8.9 | 30.6 ± 18.1 | 0.175 | |
| P | 0.008** | 0.800 | ||
| CL | Before surgery | 16.9 ± 12.2 | 15.6 ± 12.4 | 0.614 |
| 1 year after | 14.1 ± 13.0 | 7.6 ± 19.8 | 0.110 | |
| P | 0.002** | 0.014* | ||
| TS-CL | Before surgery | 10.4 ± 9.2 | 14.3 ± 9.2 | 0.048* |
| 1 year after | 11.5 ± 10.4 | 23.0 ± 17.0 | 0.000*** | |
| P | 0.167 | 0.019* | ||
| C 2-7 SVA | Before surgery | 19.8 ± 10.5 | 28.0 ± 10.3 | 0.000*** |
| 1 year after | 21.9 ± 9.8 | 44.8 ± 12.1 | 0.000*** | |
| P | 0.020* | 0.000*** | ||
| MCGS | Before surgery | −1.6 ± 9.4 | −2.3 ± 7.9 | 0.712 |
| 1 year after | −2.8 ± 7.6 | 0.8 ± 7.9 | 0.041* | |
| P | 0.169 | 0.147 | ||
| mJOA score | Before surgery | 11.2 ± 2.5 | 11.7 ± 2.8 | 0.552 |
| 1 year after | 13.3 ± 2.6 | 13.7 ± 2.6 | 0.652 | |
| P | 0.000*** | 0.029* |
*P < 0.05, **; P < 0.01, ***; P < 0.0001.
Abbreviations: NPCD; no postoperative cervical deformity; PCD; postoperative cervical deformity; TS; T1 slope, CL; cervical lordosis, C2-7 SVA; C2-7 sagittal vertical axis, MCGS; slope of McGregor’s line, mJOA score; modified Japanese orthopedic association score.
Preoperative TS-CL (14.3 ± 9.2°) and C2-7 SVA (28.0 ± 10.3 mm) in PCD group were significantly higher than those (TS-CL: 10.4 ± 9.2° and C2-7 SVA: 19.8 ± 10.5 mm in NPCD group. CL in both groups significantly reduced. However, CL in the PCD group greatly decreased from 15.6 ± 12.4° to 7.6 ± 19.8° (P = 0.014). C2-7 SVA increased significantly 1 year after surgery (NPCD group, P = 0.020; PCD group, P < 0.0001). In particular, it increased by 16.8 mm in the PCD group (P < 0.0001).
The mJOA score in NICD (11.2 ± 2.5 to 13.3 ± 2.6, P < 0.0001) and PCD (11.7 ± 2.8 to 13.7 ± 2.6, P = 0.029) groups improved significantly 1 year after surgery. However, there were no significant differences in both groups before and 1 year after surgery.
Risk Factors for Postoperative PCD, CD, and CL ≤ −10° 1 Year After Surgery
The risk factors for PCD, CD, and CL ≤ −10° 1 year after surgery were investigated using multiple logistic regression analysis (Table 5).
Table 5.
The Risk Factor of Postoperative PCD, CD, and CL ≤ −10°.
| The risk factor | Odds ratio | P value | 95% CI |
|---|---|---|---|
| Dependent variable: the presence of PCD Independent variables: spinal instability, number of decompressed lamina, decompression of C2 or C7 lamina, preoperative C2-7 SVA ≥ 30, preoperative CL ≤ 13 | |||
| Decompression of C2 or C7 vertebra | 3.10 | 0.020* | 1.19-8.03 |
| Dependent variable: the presence of CD after the surgery Independent variables: spinal instability, number of decompressed lamina, decompression of C2 or C7 vertebra, preoperative C2-7 SVA ≥ 30, preoperative CL ≤ 13 | |||
| Preoperative C2-7 SVA ≥ 30 | 19.0 | 0.000** | 8.54-42.40 |
| Dependent variable: the presence of postoperative CL ≤ −10° Independent variables: spinal instability, number of decompressed lamina, decompression of C2 or C7 vertebra, preoperative C2-7 SVA ≥ 30°, preoperative CL ≤ 2° | |||
| Preoperative CL ≤ 2° | 41.9 | 0.000** | 8.30-211.04 |
*P < 0.05, **; P < 0.0001.
Abbreviations: CD; cervical deformity, PCD; postoperative cervical deformity.
Risk Factors for PCD
The dependent variable was the presence of PCD 1 year after surgery. The independent variables were spinal instability, number of decompressed vertebrae, presence of C2 or C7 decompression, preoperative C2-7 SVA ≥ 30°, and preoperative CL ≤ 14°. The cut-off values of preoperative C2-7 SVA (≥ 30°) and CL (≤ 14°) were identified using ROC analysis. The dependent variable was presence of PCD 1 year after surgery. C2-7 SVA had a sensitivity of 59.3%, specificity of 62.7%, area under the curve (AUC) of 0.567, and 95% confidence interval (CI) of 0.472-0.663; and CL had a sensitivity of 55.4%, specificity of 44.4%, AUC of 0.502, and 95% CI of 0.387-0.617. As a result of the multiple logistic regression analysis, we found that the significant risk factor for PCD was decompression of C2 or C7 vertebra (odds ratio [OR] 3.1, P = 0.020, and 95% CI 1.19-8.03).
Risk Factors for CD
Similarly, the dependent variable was the presence of CD 1 year after the surgery, and the independent variables were spinal instability, number of decompressed vertebrae, presence of C2 or C7 decompression, preoperative C2-7 SVA ≥ 30°, and preoperative CL ≤ 13°. The result of ROC analysis was follows: C2-7 SVA (≥ 30°) had a sensitivity of 80.3%, specificity of 78.8%, AUC of 0.851, and 95% CI of 0.789-0.913 (Figure 2A); and CL (≤ 13°) had a sensitivity of 59.1%, specificity of 47.5%, AUC of 0.558, and 95% CI of 0.468-0.648. Multiple logistic regression analysis demonstrated that the significant risk factor for CD 1 year after the surgery was preoperative C2-7 SVA ≥ 30° (OR 19.0, P < 0.0001, and 95% CI 8.54-42.4).
Figure 2.
Receiver operating characteristic (ROC) analyses. A, ROC analysis to identify the cut-off value of C2-7 SVA for postoperative cervical deformity. The ROC analysis indicated a cut-off of 29.7 mm, a sensitivity of 80.3%, a specificity of 78.8%, an AUC of 0.851, and a 95% CI of 0.789-0.913. B, ROC analysis to identify the cut-off value of cervical lordosis (CL) for postoperative CL ≤ −10°. The ROC analysis indicated a cut-off of 2°, a sensitivity of 90.0%, a specificity of 69.2%, an AUC of 0.860, and a 95% CI of 0.768-0.952.
Risk Factors for Postoperative CL ≤ −10°
The dependent variable was the presence of CL ≤ −10° 1 year after surgery, and the independent variables were spinal instability, number of decompressed vertebrae, presence of C2 or C7 decompression, preoperative C2-7 SVA ≥ 30°, and preoperative CL ≤ 2°. The ROC analysis was as follows: C2-7 SVA (≥ 30°) had a sensitivity of 69.2%, specificity of 62.2%, AUC of 0.648, and 95% CI of 0.462-0.834; and CL (≤ 2°) had a sensitivity of 90.0%, specificity of 69.2%, AUC of 0.860, and 95% CI of 0.768-0.952 (Figure 2B). Finally, the significant risk factor for postoperative CL ≤ −10° was preoperative CL ≤ 2° (OR 41.9, P < 0.0001, and 95%CI 8.30-211.04).
Representative Case
A 55-year old male patient presented with gait disturbance due to spasticity, coordination disorder, and upper limb muscle weakness. He was diagnosed with cervical myelopathy due to OPLL. He was classified into the CD group because the radiographic parameters before surgery were C2-7 SVA 54 mm and CL −2° (Figure 3A). The dynamic X-ray showed no cervical instability. He underwent laminoplasty at C3 to C7. However, his cervical alignment deteriorated rapidly, and he complained of severe neck pain. Figure 3B shows the X-ray 4 months after surgery (C2-7 SVA 78 mm and CL −30°). Finally, the patient was treated surgically (C2 to T12 fusion) again half a year after the index surgery (Figure 3C). In antero-posterior view of radiograph, the upper or middle thoracic vertebrae without pedicle screw were not set because of the increased OPLL. The lower instrumented vertebra was set at T12 because it was necessary to install beyond the apical vertebra of thoracic kyphosis (T7) and the need for sufficient anchorage for long range fixation. The C2-7 SVA and CL changed to 49 mm and 0°, respectively. His neck pain also improved markedly.
Figure 3.
Representative case. A, Representative case before surgery. A 55-year-old male patient classified into the CD group. The radiographic parameters show C2-7 sagittal vertical axis (SVA) of 54 mm and cervical lordosis (CL) of −2° and no instability. B, Representative case 4 months after surgery. The patient underwent laminoplasty at C3 to C7. However, his cervical alignment deteriorated rapidly, and he complained of severe neck pain. The X-ray 4 months after surgery indicates an SVA of 5 mm, C2-7 SVA of 78 mm, and CL of −30°. C, Representative case 6 months after surgery. Finally, the patient was treated surgically (C2 to T12 fusion) due to severe neck pain. The C2-7 SVA and CL improved to 49 mm and 0°, respectively.
Discussion
Posterior decompression surgery, usually performed as laminoplasty (either open-door or double-door), is currently common as a general surgery procedure, but postoperative kyphosis remains a problem. In fact, Tables 2 and 4 demonstrated that CL decreased significantly 1 year after surgery in all groups. Postoperative CL decreased 95 of 153 patients (62.1%) in NCD group, 75 of 126 patients (59.5%) in the NPCD group, 32 of 40 patients (80%) in the CD group, and 20 of 27 patients (74.1%) in the PCD group.
It has been reported that ACSD has a strong negative impact on health-related quality of life, and hence, it is important to avoid PCD. 22 However, little is known about the causes of PCD after posterior decompression surgery, cut-off values related to its development, and course of alignment when patients with preoperative CD undergo posterior decompression. Sakai et al defined cervical kyphosis deformity as CL ≤ −5° and reported that the risk factors for iatrogenic kyphosis were center of gravity of the head—C7 SVA ≥ 42 mm and age ≥75 years. 23 On the other hand, Table 5 indicate that the risk factor for PCD found in this study was laminectomy or laminoplasty of C2 or C7. In addition, Table 3 shows that the number of decompressed vertebrae was significantly higher in the PCD group. Normally, laminoplasty involves the decompression of the 4 vertebrae of C3-6 either as direct decompression or as indirect decompression using the lordosis of the cervical spine (posterior shift of the spinal cord). In the present study, the mean decompressed vertebrae were 4 or more in all groups, and decompression of the C2 or C7 lamina in addition to the 4 vertebrae of C3-6 appeared to be a significant risk factor for PCD. In this study, PCD was found in 16 of 60 patients (26.7%) who underwent C2 or C7 lamina decompressions and who did not have CD before surgery. Therefore, if anterior compression is not severe, indirect decompression is not necessary, and it is more effective to perform selective laminoplasty and to reduce the decompressed vertebrae to prevent the development of CD. Once CD occurs, indirect decompression cannot be expected.
The incidence of C5 palsy was significantly higher in the PCD group than in the NPCD group (Table 3). The causes of C5 palsy remain unclear. Instrumentation surgery has been reported as a possible cause, 24 but this study only targeted decompression surgery. On the other hand, Imagama et al reported C5 palsy after laminoplasty was caused by stenosis of the C4/5 foramen and tethering by posterior shift of spinal cord. 25 In the PCD group, the postoperative cervical alignment was significantly altered, which may have caused narrowing of the intervertebral foramen and tethering of the spinal cord.
Preoperative C2-7 SVA ≥ 30 mm was identified as a significant risk factor for CD after the surgery (Table 5). Figure 2A indicate that the AUC, sensitivity, and specificity of this factor were all high, and the reported cut-off value seems reliable. Of the 75 patients with preoperative C2-7 SVA ≥ 30 mm, 49 (65.3%) had developed CD 1 year later. Similarly, the risk factor for postoperative CL ≤ −10° was preoperative CL ≤ 2° (Table 5). This cut-off value was also remarkably reliable in terms of AUC, sensitivity, and specificity (Figure 2B). Of the 28 patients who had CL ≤ 2° before surgery, 10 (35.7%) had CL ≤ −10° 1 year later.
Based on these results, it can be deduced that in cases where the patient presents CD before surgery, an improvement of CD after surgery is unlikely, even if posterior decompression is performed. Figure 4 shows the change in cases with and without CD before surgery 1 year after surgery. Of the 40 cases that had CD originally, only 6 (15%) did not classify for CD 1 year later.
Figure 4.
Changes in presence of cervical deformity (CD) 1 year after surgery in patients (Pt.) with and without preoperative CD. In no CD (NCD) group, 27 of 153 Pt. developed postoperative CD (PCD). The remaining 126 patients did not develop CD postoperatively (NPCD). Of the 40 cases in CD group, only 6 (15%) Pt. improved CD 1 year after surgery. On the other hand, no improvement was observed in the remaining 34 Pt.
On the other hand, CL in the CD group showed a significant decrease of 7° 1 year after surgery, but no significant change was observed in the mean value of C2-7 SVA (Table 2). In the CD group, 7 of 39 cases (17.9%) showed an increase ≥1 cm 1 year after surgery, and another 7 cases (17.9%) showed a decrease ≥1 cm. Therefore, no difference was observed in the mean values, but, as shown by the representative case, the possibility of C2-7 SVA increasing significantly in the CD group cannot be discarded. In fact, revision surgery rate for cervical malalignment in CD group was significantly higher than NCD group (Table 1).
In current study, 54 (62.1%) of 87 patients with CL ≤ 2° or C2-7 SVA ≥ 30 mm before surgery had CD 1 year after surgery. Furthermore, in the 153 patients without preoperative CD, PCD developed in 15 of 48 (31.3%) patients with CL ≤ 2° or C2-7 SVA ≥ 30 mm. From the above, it can be deduced that CL ≤ 2° and C2-7 SVA ≥ 30 mm can be useful cut-off values for postoperative CD development. The possibility of fusion surgery should be considered to avoid deterioration of postoperative CD and reoperation when the patient has preoperative CD, decompressing C2 or C7 lamina, or preoperative C2-7 SVA ≥ 30 or CL ≤ 2° is observed.
There were some limitations to this study. First, a 1-year follow-up may not be enough to evaluate changes in alignment. Second, postoperative cervical alignment can change according to the rehabilitation. But this was a retrospective multicenter study, so the details of how rehabilitation was conducted at each facility were unknown. Third, 82 of 287 patients were not included in the study because of a lack of standing radiographs; this may constitute a source of bias. Fourth, the present study did not include thoracolumbar, pelvis, or global alignment measures from whole spine standing radiographs; the study of changes in alignment using these measures needs to be addressed in the future. Fifth, postoperative outcomes were evaluated by mJOA (Tables 2 and 4), and no significant differences were observed among these groups. However, cervical malalignment is more associated with neck pain than with neurological symptoms and should be evaluated using the Neck Disability Index in future studies. 26 Sixth, 1 year of follow-up may be short to evaluate postoperative spinal alignment; studies with more than 2 years of follow-up are needed in the future.
Conclusions
The incidence of PCD was 13.9% 1 year after posterior cervical decompression surgery. CL decreased in all the groups (NCD, CD, NPCD and PCD) after surgery. In patients with preoperative CD, CD remained in 85% of the cases, even after posterior decompression surgery and the revision surgery rate was significantly higher. The risk factor for PCD was decompression of C2 or C7 lamina (OR = 3.1). At 1 year postoperatively, the risk factor for CD was preoperative C2-7 SVA ≥ 30 mm (OR = 19.0), and the risk factor for CL ≤ −10° was preoperative CL ≤ 2° (OR = 41.9). To prevent CD after the surgery, it is important to avoid decompression of the C2 or C7 lamina as much as possible. Moreover, in cases with CD, C2-7 SVA ≥ 30 mm or CL ≤ 2° before surgery, it is important to explain the risks and consider adding fusion surgery if necessary.
Footnotes
Authors’ Note: The study protocol was approved by the institutional review board of Hamamatsu University School of Medicine, Shizuoka, Japan. IRB number is 19-242.
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Shin Oe and Yu Yamato belong to the laboratory to which funds were donated, called the Division of Geriatric Musculoskeletal Health.
Funding: The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Medtronic Sofamor Danek Inc., Japan Medical Dynamic Marketing Inc., and Meitoku Medical Institution Jyuzen Memorial Hospital. We have not received funding from the NIH, HHMI, or others.
ORCID iDs: Shin Oe, MD, PhD 
https://orcid.org/0000-0002-0794-4026
Kenta Kurosu, MD
https://orcid.org/0000-0001-7660-2655
Go Yoshida, MD, PhD
https://orcid.org/0000-0003-0265-6372
Tomohiro Yamada, MD
https://orcid.org/0000-0002-7220-7321
Koichiro Ide, MD
https://orcid.org/0000-0003-4253-568X
Tatsuya Yasuda, MD, PhD
https://orcid.org/0000-0001-9587-4840
Hiroki Ushirozako, MD, PhD
https://orcid.org/0000-0002-7722-2205
References
- 1.Mikawa Y, Shikata J, Yamamuro T. Spinal deformity and instability after multilevel cervical laminectomy. Spine (Phila Pa 1976). 1987;12(1):6–11. doi:10.1097/00007632-198701000-00002 [DOI] [PubMed] [Google Scholar]
- 2.Albert TJ, Vacarro A. Postlaminectomy kyphosis. Spine (Phila Pa 1976). 1998;23(24):2738–2745. doi:10.1097/00007632-199812150-00014 [DOI] [PubMed] [Google Scholar]
- 3.Guigui P, Benoist M, Deburge A. Spinal deformity and instability after multilevel cervical laminectomy for spondylotic myelopathy. Spine (Phila Pa 1976). 1998;23(4):440–447. doi:10.1097/00007632-199802150-00006 [DOI] [PubMed] [Google Scholar]
- 4.Baisden J, Voo LM, Cusick JF, Pintar FA, Yoganandan N.Evaluation of cervical laminectomy and laminoplasty. A longitudinal study in the goat model. Spine (Phila Pa 1976). 1999;24(13):1283–1288; discussion 1288-1289. doi:10.1097/00007632-199907010-00003 [DOI] [PubMed] [Google Scholar]
- 5.Matsunaga S, Sakou T, Nakanisi K. Analysis of the cervical spine alignment following laminoplasty and laminectomy. Spinal Cord. 1999; 37(1):20–24. doi:10.1038/sj.sc.3100749 [DOI] [PubMed] [Google Scholar]
- 6.Ratliff JK, Cooper PR. Cervical laminoplasty: a critical review. J Neurosurg. 2003;98(3 suppl):230–238. doi:10.3171/spi.2003.98.3.0230 [DOI] [PubMed] [Google Scholar]
- 7.Miyazaki K, Tada K, Matsuda Y, Okuno MA, Yasuda TA, Murakami HI. Posterior extensive simultaneous multisegment decompression with posterolateral fusion for cervical myelopathy with cervical instability and kyphotic and/or S-shaped deformities. Spine (Phila Pa 1976). 1989;14(11):1160–1170. doi:10.1097/00007632-198911000-00006 [DOI] [PubMed] [Google Scholar]
- 8.Herkowitz HN. A comparison of anterior cervical fusion, cervical laminectomy, and cervical laminoplasty for the surgical management of multiple level spondylotic radiculopathy. Spine (Phila Pa 1976). 1988;13(7):774–780. doi:10.1097/00007632-198807000-00011 [DOI] [PubMed] [Google Scholar]
- 9.Hamanishi C, Tanaka S. Bilateral multilevel laminectomy with or without posterolateral fusion for cervical spondylotic myelopathy: relationship to type of onset and time until operation. J Neurosurg. 1996;85(3):447–451. doi:10.3171/jns.1996.85.3.0447 [DOI] [PubMed] [Google Scholar]
- 10.Hirabayashi K, Watanabe K, Wakano K, Suzuki N, Satomi K, Ishii Y. Expansive open-door laminoplasty for cervical spinal stenotic myelopathy. Spine (Phila Pa 1976). 1983;8(7):693–699. doi:10.1097/00007632-198310000-00003 [DOI] [PubMed] [Google Scholar]
- 11.Kawai S, Sunago K, Doi K, Saika M, Taguchi T.Cervical laminoplasty (Hattori’s method). Procedure and follow-up results. Spine (Phila Pa 1976). 1988;13(11):1245–1250. PMID: 3144758 [PubMed] [Google Scholar]
- 12.Yukawa Y, Kato F, Ito K, et al. Laminoplasty and skip laminectomy for cervical compressive myelopathy: range of motion, postoperative neck pain, and surgical outcomes in a randomized prospective study. Spine (Phila Pa 1976). 2007;32(18):1980–1985. doi:10.1097/BRS.0b013e318133fbce [DOI] [PubMed] [Google Scholar]
- 13.Nakashima H, Kato F, Yukawa Y, et al. Comparative effectiveness of open-door laminoplasty versus French-door laminoplasty in cervical compressive myelopathy. Spine (Phila Pa 1976). 2014;39(8):642–647. doi:10.1097/BRS.0000000000000252 [DOI] [PubMed] [Google Scholar]
- 14.Miyazaki M, Ishihara T, Notani N, Kanezaki S, Tsumura H. Relationship of T1 slope with loss of lordosis and surgical outcomes after laminoplasty for cervical ossification of the posterior longitudinal ligament. Clin Neurol Neurosurg. 2018;164:19–24. doi:10.1016/j.clineuro.2017.11.007 [DOI] [PubMed] [Google Scholar]
- 15.Uchida K, Nakajima H, Sato R, et al. Cervical spondylotic myelopathy associated with kyphosis or sagittal sigmoid alignment: outcome after anterior or posterior decompression. J Neurosurg Spine. 2009;11(5):521–528. doi:10.3171/2009.2.SPINE08385 [DOI] [PubMed] [Google Scholar]
- 16.Fehlings MG, Wilson JR, Kopjar B, et al. 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. 2013;95(18):1651–1658. doi:10.2106/JBJS.L.00589 [DOI] [PubMed] [Google Scholar]
- 17.Chavanne A, Pettigrew DB, Holtz JR, Dollin N, Kunt C, IV. Spinal cord intramedullary pressure in cervical kyphotic deformity: a cadaveric study. Spine (Phila Pa 1976). 2011;36(20):1619–1626. doi:10.1097/BRS.0b013e3181fc17b0 [DOI] [PubMed] [Google Scholar]
- 18.Tang JA, Scheer JK, Smith JS, et al. ISSG. The impact of standing regional cervical sagittal alignment on outcomes in posterior cervical fusion surgery. Neurosurgery. 2012;71(3):662–669; discussion 669. doi:10.1227/NEU.0b013e31826100c9 [DOI] [PubMed] [Google Scholar]
- 19.Oe S, Togawa D, Nakai K, et al. The influence of age and sex on cervical spinal alignment among volunteers aged over 50. Spine (Phila Pa 1976). 2015;40(19):1487–1494. doi:10.1097/BRS.0000000000001071 [DOI] [PubMed] [Google Scholar]
- 20.Oe S, Togawa D, Yoshida G, et al. Difference in spinal sagittal alignment and health-related quality of life between males and females with cervical deformity. Asian Spine J. 2017;11(6):959–967. doi:10.4184/asj.2017.11.6.959 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Suda K, Abumi K, Ito M, Shono Y, Kaneda K, Fujiya M. Local kyphosis reduces surgical outcomes of expansive open-door laminoplasty for cervical spondylotic myelopathy. Spine (Phila Pa 1976). 2003;28(12):1258–1262. doi:10.1097/01.BRS.0000065487.82469.D9 [DOI] [PubMed] [Google Scholar]
- 22.Smith JS, Line B, Bess S, et al. Group ISS. The health impact of adult cervical deformity in patients presenting for surgical treatment: comparison to United States population norms and chronic disease states based on the EuroQuol-5 dimensions questionnaire. Neurosurgery. 2017;80(5):716–725. doi:10.1093/neuros/nyx028 [DOI] [PubMed] [Google Scholar]
- 23.Sakai K, Yoshii T, Hirai T, et al. Cervical sagittal imbalance is a predictor of kyphotic deformity after laminoplasty in cervical spondylotic myelopathy patients without preoperative kyphotic alignment. Spine (Phila Pa 1976). 2016;41(4):299–305. doi:10.1097/BRS.0000000000001206 [DOI] [PubMed] [Google Scholar]
- 24.Takemitsu M, Cheung KM, Wong YW, Cheung WY, Luk KD. C5 nerve root palsy after cervical laminoplasty and posterior fusion with instrumentation. J Spinal Disord Tech. 2008;21(4):267–272. doi:10.1097/BSD.0b013e31812f6f54 [DOI] [PubMed] [Google Scholar]
- 25.Imagama S, Matsuyama Y, Yukawa Y, et al. Nagoya Spine Group. C5 palsy after cervical laminoplasty: a multicentre study. J Bone Joint Surg Br. 2010;92(3):393–400. doi:10.1302/0301-620X.92B3.22786 [DOI] [PubMed] [Google Scholar]
- 26.Oe S, Togawa D, Yoshida G, et al. Cut-off values of and factors associated with a negative influence on the neck disability index. Eur Spine J. 2018;27(6):1423–1431. doi:10.1007/s00586-018-5555-7 [DOI] [PubMed] [Google Scholar]