Abstract
Study DesignRetrospective cohort study.ObjectivesTo determine whether preoperative cervical bone quality assessed by computed tomography Hounsfield units (CT HU) predicts pseudarthrosis, cage subsidence, or loss of cervical lordosis after anterior cervical discectomy and fusion (ACDF) and to identify perioperative predictors.MethodsConsecutive patients who underwent 1-3-level ACDF with ≥5-year follow-up were reviewed. Outcomes included neck/arm pain visual analog scale scores, Neck Disability Index, and C2-C7 lordosis and segmental height. Bone quality was assessed by preoperative CT HU. Subgroup comparisons, multivariable regression, and receiver operating characteristic (ROC) analyses evaluated predictors.ResultsAmong 253 patients, subsidence, pseudarthrosis, and loss of lordosis occurred in 15 (5.9%), 26 (10.3%), and 41 (16.2%), respectively. Index HU and clinical outcomes did not differ between subgroups defined by subsidence or pseudarthrosis. HU ≤ 231.5 was associated with a higher incidence of loss of lordosis. On regression, higher BMI and greater segmental lengthening independently predicted subsidence, while multilevel surgery predicted pseudarthrosis. Greater preoperative C2-C7 lordosis predicted loss of lordosis; very low HU (≤231.5) was significant in univariable analysis but not after multivariable adjustment. On ROC analysis, Index HU did not predict subsidence (AUC: 0.511), pseudarthrosis (AUC: 0.414), or loss of lordosis (AUC: 0.551). In contrast, segmental lengthening predicted subsidence (AUC 0.695) with threshold around 3 mm (3.12 mm).ConclusionPreoperative cervical CT HU did not predict pseudarthrosis or subsidence over ≥5 years after ACDF; low HU showed an unadjusted association with loss of lordosis. Bone quality alone may be insufficient to justify avoiding ACDF. Limiting segmental lengthening to ∼3 mm may mitigate subsidence risk.