CT Hounsfield Units as a predictor for the worsening of traumatic vertebral compression fractures

Introduction With vertebral compression fractures (VCF), height loss has been associated with kyphotic deformity and intractable pain—two indications for potential surgical intervention. Consequently, assessment of factors associated with continual height loss can provide insights regarding management. Computed tomography Hounsfield units (CT HU), a measure of radio-density, has been implicated for the assessment of bone quality. No studies have assessed the relationship between CT HU and traumatic VCF. Consequently, the objective of this study was to evaluate this relationship. Methods Patients who sustained a traumatic VCF from a motor vehicle accident from 2006 to 2015 were reviewed retrospectively. Inclusion criteria were single VCF from L1 to L4, appropriate imaging, and appropriate follow up. Clinical data were extracted [age, gender, history of osteoporosis, body mass index (BMI), and length of follow-up]. Initial CT imaging was evaluated for averaged HU at the levels above and below the index level of the VCF. Interval percentage change (Δ%) in anterior compression was evaluated between x rays obtained during initial presentation and at follow-up. Results Age, BMI, and CT HU significantly correlated with Δ% (p < 0.01) while initial height loss, history of osteoporosis, and length of follow up did not. A multiple linear regression for Δ% that combined age, BMI, and CT HU values elicited a regression model with R2 0.648, and adjusted R2 0.606; each variable in the model exhibited p < 0.05. An ROC analysis to predict Δ% < 10% from Δ% ≥10% revealed an area under the curve 0.983, p < 0.01, where CT HU at 171.58 exhibited a combination of sensitivity at 0.923 and specificity at 0.91. Conclusion Age, BMI, and CT HU values are independent predictors of worsening VCF. These factors can help determine appropriate clinical follow-up and need for surgical intervention.