Caudally directed upper-instrumented vertebra pedicle screws associated with minimized risk of proximal junctional failure in patients with long posterior spinal fusion for adult spinal deformity

Abstract

BACKGROUND CONTEXTIt is unknown whether upper instrumented vertebra (UIV) pedicle screw trajectory and UIV screw-rod angle are associated with development of proximal junctional kyphosis (PJK) and/or proximal junctional failure (PJF). PURPOSETo determine whether (1) the cranial-caudal trajectory of UIV pedicle screws and (2) UIV screw-vertebra angle are associated with PJK and/or PJF after long posterior spinal fusion in patients with adult spinal deformity (ASD). STUDY DESIGN/SETTINGRetrospective review. PATIENT SAMPLEWe included 96 patients with ASD who underwent fusion from T9–T12 to the pelvis (>5 vertebrae fused) between 2008 and 2015. OUTCOME MEASURESPedicle screw trajectory was measured as the UIV pedicle screw–vertebra angle (UIV-PVA), which is the mean of the two angles between the UIV superior endplate and both UIV pedicle screws. (Positive values indicate screws angled cranially; negative values indicate screws angled caudally.) We measured UIV rod-vertebra angle (UIV-RVA) between the rod at the point of screw attachment and the UIV superior endplate. METHODSDuring ≥2-year follow-up, 38 patients developed PJK, and 28 developed PJF. Mean (± standard deviation) UIV-PVA was −0.9° ± 6.0°. Mean UIV-RVA was 87° ± 5.2°. We examined the development of PJK and PJF using a UIV-PVA/UIV-RVA cutoff of 3° identified by a receiver operating characteristic curve, while controlling for osteoporosis, age, sex, and preoperative thoracic kyphosis. RESULTSPatients with UIV-PVA ≥3° had significantly greater odds of developing PJK (odds ratio 2.7; 95% confidence interval: 1.0–7.1) and PJF (odds ratio 3.6; 95% confidence interval: 1.3–10) compared with patients with UIV-PVA <3°. UIV-RVA was not significantly associated with development of PJK or PJF. CONCLUSIONSIn long thoracic fusion to the pelvis for ASD, UIV-PVA ≥3° was associated with 2.7-fold greater odds of PJK and 3.6-fold greater odds of PJF compared with UIV-PVA <3°. UIV-RVA was not associated with PJK or PJF. LEVEL OF EVIDENCEIII