The impact of baseline cervical malalignment on the development of proximal junctional kyphosis following surgical correction of thoracolumbar adult spinal deformity.

Document Type


Publication Date


Publication Title

Journal of neurosurgery. Spine


washington; seattle; swedish neuro


OBJECTIVE: The objective of this study was to identify the effect of baseline cervical deformity (CD) on proximal junctional kyphosis (PJK) and proximal junctional failure (PJF) in patients with adult spinal deformity (ASD).

METHODS: This study was a retrospective analysis of a prospectively collected, multicenter database comprising ASD patients enrolled at 13 participating centers from 2009 to 2018. Included were ASD patients aged > 18 years with concurrent CD (C2-7 kyphosis < -15°, T1S minus cervical lordosis > 35°, C2-7 sagittal vertical axis > 4 cm, chin-brow vertical angle > 25°, McGregor's slope > 20°, or C2-T1 kyphosis > 15° across any three vertebrae) who underwent surgery. Patients were grouped according to four deformity classification schemes: Ames and Passias CD modifiers, sagittal morphotypes as described by Kim et al., and the head versus trunk balance system proposed by Mizutani et al. Mean comparison tests and multivariable binary logistic regression analyses were performed to assess the impact of these deformity classifications on PJK and PJF rates up to 3 years following surgery.

RESULTS: A total of 712 patients with concurrent ASD and CD met the inclusion criteria (mean age 61.7 years, 71% female, mean BMI 28.2 kg/m2, and mean Charlson Comorbidity Index 1.90) and underwent surgery (mean number of levels fused 10.1, mean estimated blood loss 1542 mL, and mean operative time 365 minutes; 70% underwent osteotomy). By approach, 59% of the patients underwent a posterior-only approach and 41% underwent a combined approach. Overall, 277 patients (39.1%) had PJK by 1 year postoperatively, and an additional 189 patients (26.7%) developed PJK by 3 years postoperatively. Overall, 65 patients (9.2%) had PJF by 3 years postoperatively. Patients classified as having a cervicothoracic deformity morphotype had higher rates of early PJK than flat neck deformity and cervicothoracic deformity patients (p = 0.020). Compared with the head-balanced patients, trunk-balanced patients had higher rates of PJK and PJF (both p < 0.05). Examining Ames modifier severity showed that patients with moderate and severe deformity by the horizontal gaze modifier had higher rates of PJK (p < 0.001).

CONCLUSIONS: In patients with concurrent cervical and thoracolumbar deformities undergoing isolated thoracolumbar correction, the use of CD classifications allows for preoperative assessment of the potential for PJK and PJF that may aid in determining the correction of extending fusion levels.

Clinical Institute

Orthopedics & Sports Medicine

Clinical Institute

Neurosciences (Brain & Spine)