Objective:To compare the clinical outcomes and correction effects of kyphosis between Zero-profile device (Zero-p) and plate/cage structures (PCC) in treating cervical spondylotic myelopathy (CSM) patients with cervical kyphosis.Methods:From August 2016 to July 2018, a total of 54 cases of cervical spondylotic myelopathy patients with cervical kyphosis were analyzed retrospectively, including 26 cases treated with Zero-p and 28 cases treated with PCC system. There was no significant difference between the two groups in gender, age, body mass index (BMI) and operative segment. The operation duration and the blood loss were recorded. The clinical outcomes of the patients were measured by visual analogue score (VAS) for neck pain and Japanese Orthopedic Association (JOA) score for neurological function. Moreover, JOA recovery rate was obtained to assess the surgical results. The cervical lordosis (C 2-C 7 Cobb angle), the Cobb angle of the operation segment, the C 2-C 7 vertical axis (C 2 SVA) and the cervical range of motion (ROM) were measured on the lateral and dynamic radiographs of the cervical spine, respectively. Results:In the Zero-p group, the operation duration was 83.0±14.9 (range 60-120) min, intraoperative blood loss was 70.5±27.3 (range 30-150) ml. In PCC group, the operation duration was 100.0±23.9 (range 65-145) min, intraoperative blood loss was 104.2±38.8 (range 30-250) ml. There were significant difference in above parameters between two groups ( t=3.40, 2.06; P=0.00, 0.04). The follow-up duration in Zero-p group was 30.4±5.8 (range 24-36) months and 31.2±4.9 (range 24-36) months in PCC group without significant difference ( t=1.061, P=0.291). The VAS/JOA score of the Zero-p group was improved from (5.9±1.0)/(9.2±1.7) preoperatively to (2.1±0.8)/(14.9±1.0) at 1 month postoperatively, and to (3.4±1.0)/(15.1±0.9) at the last follow-up. The difference between them was statistically significant ( F=130.96, 221.40, P=0.00). The VAS/JOA score of the PCC group was improved from (5.9±1.1)/(8.7±1.6) preoperatively to (2.3±0.9)/(14.9±1.0) at 1 month after surgery, and to (2.6±0.9)/(15.6±1.1) at the last follow-up. The difference between them was statistically significant ( F=303.35, 126.64, P=0.00). However, the VAS score of neck pain in the Zero-p group at the last follow-up was significantly deteriorated, which was significantly higher than that in PCC group ( P<0.05). The cervical lordosis/operative segment Cobb angle in the Zero-p group was improved from preoperative (-6.7°±2.7°)/(-6.5°±3.2°) preoperatively to (14.2°±4.9°)/(12.9°±4.9°) at 1 month postoperatively, and to (5.9°±4.7°)/(5.0°±4.0°) at the last follow-up with statistical significance ( F=196.98, 179.97, P=0.00). The cervical lordosis/operative segment Cobb angle in the PCC group was improved from (-5.7°±3.5°)/(-6.1°±4.0°) preoperatively to (13.9°±6.9°)/(13.0°±6.4°) 1 month after surgery, and to (11.0°±5.5°)/(10.4°±5.6°) at the last follow-up with statistical significance ( F=127.27, 119.98, P=0.00). However, the cervical lordosis and operative segment Cobb angle at the last follow-up in the Zero-p group were significantly lost compared with those at 1 month after surgery, which were significantly smaller than those in the PCC group ( P<0.05). The incidence of dysphagia after operation was 7.7% (2/26) in the Zero-p group and 28.6% (8/28) in the PCC group (χ 2=5.11, P=0.02). Conclusion:For CSM patients with cervical kyphosis, PCC could achieve much better mid-term kyphotic correction and clinical outcomes. However, Zero-p should be avoided as much as possible.