Influence of complete uncinate process removal on 2-year assessment of radiologic outcomes: subsidence and sagittal balance in patients receiving one-level anterior cervical discectomy and fusion

Background: Many patients with cervical radiculopathy experience stenosis of the neural foramens due to cumulative osteophyte or uncovertebral joint hypertrophy. For cervical foraminal stenosis, complete uncinate process resection (UPR) is often conducted concurrently with anterior discectomy and fusion (ACDF). The aim of this study was to evaluate the clinical and radiological consequences of complete UPR during ACDF versus those seen with ACDF without UPR. Methods: In total, 105 patients who underwent one-level ACDF with a cage-and-plate construct between 2011 and 2015 were retrospectively reviewed. Among them, 37 underwent ACDF with complete UPR, and 68 underwent ACDF without UPR. Radiographic parameters of disc height, C2–C7 lordosis, T1 slope, C2–C7 sagittal vertical axis (SVA), center of the sella turcica–C7 SVA (St-SVA), spino-cranial angle (SCA), and fusion rate were measured on plain radiographs at pre-operation, immediately post-operation, and during the follow-up period (median follow-up duration: 37.7 ± 10.5 months). Results: All of the clinical parameters improved at the 2-year follow up (P<0.0001). Improvement in visual analogue scale (VAS) scores for arm pain was significantly better in the ACDF with complete UPR group immediately post-operation. All cervical sagittal parameters, including cervical lordosis, segmental angle, disc height, C2-C7 SVA, St-SVA, T1 slope, and SCA, were similar between the ACDF with UPR and ACDF without UPR groups. Differences in segmental angle, disc height, C2-C7 SVA, St-SVA, and SVA at 2-year follow up after preoperative examination, however, were statistically significant (p<0.05). Subsidence occurred in 23 patients (ACDF with complete UPR: 14 cases 37% versus ACDF without UPR: 9 cases 13%; p < 0.05). Conclusions: Cervical sagittal alignment after ACDF with complete UPR is not significantly different from that achieved with ACDF

occur more frequently after ACDF with complete UPR than after ACDF without UPR, although with little to no clinical impact. More precise and careful selection of patients is needed when deciding on additional complete UPR.

Background
Anterior cervical discectomy and fusion (ACDF) aiming to improve the stability of the vertebra by decompression of neural elements and fusion is regarded as the gold-standard procedure for symptomatic cervical spondylosis in patients in whom non-operative care has failed [1]. Clinical and radiologic results after ACDF appear to be good [2]. Many patients with cervical radiculopathy also experience stenosis of the neural foramens because of cumulative osteophyte or uncovertebral joint hypertrophy. Although most anterior cervical discectomy and fusion procedures include cervical uncosectomy or uncoforaminotomy to decompress nerve roots in patients with cervical radiculopathy, Lee DH et al. reported that complete uncinate process resection (UPR) during ACDF improves pain in a patient's arm more rapidly than conventional ACDF without UPR and provides similar fusion rates [3,5]. Meanwhile, SH Lee et al. reported that complete UPR over 38% during ACDF increases the risk of subsidence during follow up [4].
At present, there is little evidence of whether this surgical technique provides good clinical and radiologic outcomes after complete unilateral or bilateral UPR, especially in regards to subsidence and cervical sagittal alignment. Accordingly, this study was undertaken to evaluate the influence of complete UPR on subsidence and regional cervical sagittal balance by comparing the clinical and radiologic outcomes after ACDF with complete UPR versus ACDF without UPR.

Patient recruitment and inclusion criteria
Between January 2011 and December 2015, 578 patients who underwent ACDF for cervical spondylotic disease at our institution were collected. Among them, we excluded 473 patients whose follow-up period was less than 2 years or the surgery level was two levels or more. In this retrospective study, 105 consecutive patients with single-level cervical spondylotic disease who underwent primary ACDF with a cage-and-plate construct between January 2011 and December 2015 at the author's institution were included (Fig.   1). This study was approved by the Institutional Review Board of our hospital. The uncinate process was randomly removed totally according to the technical preference of the single surgeon (Fig. 2). Thus, we defined ACDF with UPR as complete unilateral or bilateral removal of the uncinate process, while ACDF without UPR was defined as the conventional removal of only the anterior and posterior parts of the uncinate process or no removal of the uncinate process. This was confirmed with postoperative computed tomography scans. The patients were divided into two groups: 37 patients underwent ACDF with complete UPR and 68 patients were treated with ACDF without UPR. The inclusion criteria included the following: 1) patients with symptoms of degenerative cervical disease; 2) patients who received primary ACDF with UPR at only one level; and 3) a follow-up period greater than 24 months. The exclusion criteria were as follows: 1) patients who had previous cervical spine surgery due to ossification of posterior longitudinal ligaments, fractures, tumors, etc.; 2) patients who underwent ACDF for more than two levels; and 3) a follow-up period less than 24 months.

Surgical procedure
The patients were positioned under general anesthesia in the supine position. The surgical technique was chosen using a standard Smith-Robinson technique. After confirmation and exposure of the proper vertebral levels according to the compressive materials, a discectomy was performed, and a high-speed burr was applied to remove the anterior and posterior bony spurs and the endplate cartilage. The endplate cartilage was eliminated with a curette carefully to preserve the bony endplate as much as possible to prevent cage subsidence. Discs, endplate cartilaginous, and other compressive materials were subducted to achieve appropriate dural and neural decompression. Using an osteotome, a high-speed electric drill, and a Kerrison punch, the nerve roots were decompressed by completely removing the uncinate process. If the patient had unilateral symptoms and if radiologic results were consistent, we performed removal of the uncinate process unilaterally. We used a plate (Atlantis; Medtronic, Minneapolis, MN, USA) and allograft cage (Cornerstone®-SR; Medtronic, Minneapolis, MN, USA) with local autologous bone. We did not use autologous iliac bone or growth factors, such as demineralized bone matrix and recombinant bone morphogenetic proteins (rhBMP), as graft material. The proper size for the allobone cage was decided by both preoperative evaluation and intraoperative formatting using a trial cage. The cage was placed into the disc space as described above. Fixed type screws were utilized to fix the anterior cervical plate. If there was no complication during operation, all patients were able to sit upright and walk with a neck collar on the first day after surgery. The patients wore a cervical collar for 1 month after surgery. Clinical and radiographic results were obtained by an independent observer for 5 days post-operatively. In the outpatient clinic, patients were continuously followed up post-operation.

Clinical outcome assessment
Intraoperative blood loss, operative time, days of hospitalization, and clinical outcomes were evaluated using the neck disability index (NDI), neck visual analog scale (VAS), and arm-VAS preoperatively immediately after operation and at 2-year follow up. During the last follow up, the patient was assessed according to Odom's criteria, from poor to excellent [6].

Radiological evaluation
Preoperative radiologic examination evaluated plain radiographs, computed tomography scans, and magnetic resonance imaging. Plain radiological examinations of the cervical spine were also conducted immediately after surgery and at 2-year follow up for all patients. Cervical alignment was evaluated using the Cobb angle of C2-C7, working the process described by Borden [7]: this angle was made by the lines along the inferior  (Fig. 3D). We decided the maximum difference in the O-s values at each examination as 2 degrees. Radiological fusion was decided to have occurred when there was ≤ 2° movement on flexion-extension and/or ≤ 2 mm of movement of the interspinous distance on flexion-extension across the fusion segment [9].

Statistical analysis
The findings are presented as mean values ± standard deviations (SD) or counts, as indicated. The independent t-test and chi-squared test results were used to compare both groups. The multivariable logistic regression test was used to determine the influencing radiologic factors of subsidence. All P values <0.05 were considered to indicate statistical significance. All statistical analyses were performed using SPSS (version 23.0, SPSS, Chicago, IL, USA).
Regarding cervical alignment in the two groups, segmental angles at 2-year follow up were markedly better than those preoperatively (p<0.05 were only a few cases of removal of uncinate on both sides. However, subsidence occurred more frequently in cases of removal of both sides than in cases of removing only one side.

4.
Multivariate analysis of the five measurements as significant parameters on subsidence ( Table 4) Radiologic factors that may potentially influence subsidence were analyzed using logistic regression test. The results are shown in Table 4. As an influencing factor of subsidence, preoperative C2-7 SVA, St-SVA, and SCA values were significant (P<0.05). In opposition to our hypothesis, complete UPR was not a significant factor affecting subsidence.

Discussion
ACDF is the treatment of choice for symptomatic cervical spondylosis in patients when conservative treatments, such as medication or physiotherapy, have failed [10]. Patients with arm pain with neural foramen stenosis due to osteophytes or hypertrophy of the uncovertebral joint should be treated with ACDF, as well as UPR. ACDF with complete UPR is known to improve pain in the arm better and faster [11]. However, inadequate removal of the uncinate process has been reported to contribute to poor outcomes in cervical spondylosis cases [12]. In our study, the ACDF with UPR group had better arm pain in the immediate post-operation period than the ACDF without UPR group.
As the uncinate process is an important structure to maintaining the stability of adjacent vertebral bodies in the spinal axis, we investigated whether sagittal alignment or subsidence is affected by removing the uncinate process. Subsidence occurs as a natural process during the course of an interbody fusion procedure and is described as settlement of a body with a higher elasticity modulus (e.g., graft, cage, spacer) into a body with lower elasticity modulus (e.g., vertebral body), leading to a change in spine structure [13]. However, upon excessive subsidence, interbody spaces are narrowed and kyphosis of the spine occurs. This introduces instability of the screw-plate and screw-bone (e.g., pullout, change of angulation, breakage of the instrumentation) [13]. To the best of our knowledge, end-plate preparation, type of cage and size, multilevel fusion, recombinant human bone morphogenetic protein-2 (rhBMP-2), process of instrumentation, and bone quality are significant factors of subsidence [14]. In our study, when the ACDF with complete UPR and ACDF without UPR were compared under the same conditions, subsidence was significantly higher when complete UPR was performed after 3 years on average. Considering these reasons, it would seem that end-plate preparations would be performed more in the process of UPR in the ACDF with UPR group. However, between the ACDF with UPR and ACDF without UPR groups, clinical outcomes were not significantly different. This is because the foramen is widened due to the UPR, such that, even if subsidence occurs, radiculopathy due to pressing of the root does not occur. Overall, in the case of one-level ACDF, it is difficult to find a significant adverse effect of subsidence.
However, caution against subsidence is needed, and a large-scale and long-term follow-up study of multiple-level ACDF with UPR is necessary.
Sagittal balance has been suggested for cervical spine treatment. T1 slope determines the sagittal balance of the cervical spine, and this parameter is related with C2-C7 angle [15]. Previous studies have reported that C2-C7 lordosis is closely related to the other cervical and thoracic parameters (cervical lordosis, thoracic kyphosis) [16]. Cervical sagittal imbalance influences the health-related quality of life (HRQOL) of patients [17].
St-SVA and C2-C7 SVA are closely associated with the clinical results of neck pain and HRQOL [18]. The A study by Tang et al. suggested that increasing cervical SVA is a cause for clinical concern of cervical malalignment as reflected by poor HRQOL scores [19]. In our study, C2-C7 lordosis, segmental angle, disc height, C2-C7 SVA, St-SVA, T1 slope, and SCA were not different between ACDF with UPR and ACDF without UPR group, although the differences significant in segmental angle, disc height, C2-C7 SVA, St-SVA, and SVA at last follow-up and preoperatively were statistically between the two surgery groups (p<0.05).
Accordingly, there were no differences in clinical outcomes between the two groups.
Global cervical spine lordosis was not influenced by single-level ACDF [20]. This is the natural mechanism of the human body, which keeps the head on a neutral axis in the optimal horizontal plane for the visiovestibular system and re-establishes sagittal balance [20]. In our study, single-level ACDF with UPR did not affect sagittal balance, although parameters of segmental angle, disc height, C2-C7 SVA, St-SVA, and SVA were worse.
Thus, long-term follow up and a large scale study of multiple-level ACDF with UPR or ACDF in kyphotic cervical spine are necessary. Technically, UPR usually proceeds from the inside to the outside. This technique needs to be performed carefully because of the possibility of injury to the nerve roots and vertebral arteries. It is recommended to use a punch rather than a drill when removing the lateral portion of the uncinated process.

Limitations of this study
Our study had a few limitations. The number of patients who underwent removal of the uncinate process was small. Also, cases with a bilaterally UPR were rare. Also, because our study did not have a randomized controlled design, we could not completely control the possibility of selection bias. Additionally, because our study size was small, we were limited in our ability to make comparisons between the groups for several factors known to affect prognosis. Failure to indicate the extent to which the uncinate process was removed as an objective indicator was also a limitation. However, the results of this study suggest that when performing ACDF with complete UPR, the risk of subsidence should be considered. Prospective studies will be conducted using well-guided evidence-based protocols with adequate controls.

Conclusions
Cervical sagittal alignment after ACDF with complete UPR is not significantly different from that achieved with ACDF without UPR. However, subsidence occurs more frequently after ACDF with complete UPR than after ACDF without UPR, although this appears to have no clinical impact. More precise and careful selection of patients is needed when deciding on additional complete UPR.  Figure 1 Flow chart of the patients in our study.   A case from the ACDF with complete UPR group. The patient underwent an ACDF operation of C5/6 with complete UPR. In this patient, C2-C7 SVA and St-SVA increased with time, but SCA decreased with time.

Figure 5
A case from the ACDF without UPR group. The patient underwent an ACDF operation of C4/5 without UPR. In this patient, C2-C7 SVA and St-SVA decreased with time, but SCA increased with time.